Bypass reconstruction method and computer equipment

Abstract

The present application relates to a bypass reconstruction method and computer device. The method includes: obtaining heart segmentation results and blood vessel segmentation results, obtaining a target blood vessel mask based on the heart segmentation results, the blood vessel segmentation results, and a segmentation model, obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation results, and processing the bypass reconstruction information to obtain a bypass reconstruction result. This method can obtain a target blood vessel mask based on a neural network model, further obtain bypass reconstruction information based on the target blood vessel mask, and then reconstruct and determine the bypass reconstruction result based on the bypass reconstruction information. This method can avoid manual reconstruction of the bypass model, save human resources and bypass reconstruction time, reduce bypass reconstruction costs, and further improve bypass reconstruction efficiency.

Description

Technical Field

[0001] The present application relates to the field of medical technology, and in particular to a bypass reconstruction method and computer equipment. Background Art

[0002] Cardiovascular disease has become a common illness, with high morbidity and mortality rates. Coronary heart disease, in particular, is particularly severe, significantly impacting people's daily lives. Coronary artery bypass grafting (CABG) is a commonly used and most effective surgical procedure for treating CHD, making regular follow-up of patients after CABG surgery essential.

[0003] Traditionally, bypass data is manually reconstructed after surgery to help doctors understand the results. However, this traditional method can lead to low bypass reconstruction efficiency. Summary of the Invention

[0004] Based on this, it is necessary to provide a bridge reconstruction method and computer equipment to address the above technical problems.

[0005] A bridge reconstruction method, comprising:

[0006] Obtain heart segmentation results and blood vessel segmentation results;

[0007] Acquire a target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result, and the segmentation model;

[0008] Obtaining bypass reconstruction information based on the target vessel mask;

[0009] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0010] A bridge reconstruction method, comprising:

[0011] Obtain heart segmentation results and blood vessel segmentation results;

[0012] Acquire a target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result, and the segmentation model;

[0013] Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result, the bypass reconstruction information including a bypass starting point and an anastomosis between the bypass and the normal blood vessel;

[0014] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0015] A bridge reconstruction method, comprising:

[0016] Obtain heart segmentation results and blood vessel segmentation results;

[0017] Acquire a target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result, and the segmentation model;

[0018] Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result, the bypass reconstruction information including a bypass trajectory and an anastomosis between the bypass and the normal blood vessel;

[0019] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0020] A computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0021] Obtain heart segmentation results and blood vessel segmentation results;

[0022] Acquire a target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result, and the segmentation model;

[0023] Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result, the bypass reconstruction information including a bypass trajectory and an anastomosis between the bypass and the normal blood vessel;

[0024] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0025] The above-mentioned bypass reconstruction method and computer equipment, the computer equipment can obtain the heart segmentation results and the blood vessel segmentation results, obtain the target blood vessel mask based on the heart segmentation results, the blood vessel segmentation results and the segmentation model, obtain the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation results, and obtain the bypass reconstruction result by processing the bypass reconstruction information; the above-mentioned method obtains the bypass reconstruction information based on the neural network model, and determines the bypass reconstruction result by processing the bypass reconstruction information, which can avoid manual reconstruction of the bypass model, save human resources and bypass reconstruction time, reduce the bypass reconstruction cost, and further improve the bypass reconstruction efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a diagram of the internal structure of a computer device in one embodiment;

[0027] Figure 2 Schematic diagram of a flow chart of a bridge reconstruction method in one embodiment;

[0028] Figure 3 A schematic flow chart of a method for obtaining a target blood vessel mask according to one embodiment;

[0029] Figure 4 A schematic flow chart of a method for obtaining an initial blood vessel mask in another embodiment;

[0030] Figure 5 A schematic flow chart of a specific method for obtaining a target blood vessel mask in another embodiment;

[0031] Figure 6 is a flowchart of a method for determining whether a first blood vessel segmentation result includes bypass data in another embodiment;

[0032] Figure 7 This is a flowchart of a specific method for determining whether a first blood vessel segmentation result includes bypass data in another embodiment;

[0033] Figure 7a A schematic diagram of a process result of performing a bypass reconstruction method on a cardiac image of a diagnosis subject in another embodiment;

[0034] Figure 8 Schematic diagram of a flow chart of a bridge reconstruction method in one embodiment;

[0035] Figure 9 A schematic flow chart of a method for obtaining bridge reconstruction information in one embodiment;

[0036] Figure 10 A schematic flow chart of a method for determining candidate bridge points in one embodiment;

[0037] Figure 11 A schematic flow chart of a method for determining a bridge starting point in another embodiment;

[0038] Figure 12 A schematic flow chart of a specific method for determining a bridge starting point in another embodiment;

[0039] Figure 13 A schematic flow chart of a specific method for obtaining bridge reconstruction information in another embodiment;

[0040] Figure 14 Schematic diagram of a flow chart of a bridge reconstruction method in one embodiment;

[0041] Figure 15 A schematic flow chart of a method for obtaining bridge reconstruction information in one embodiment;

[0042] Figure 16 A schematic flow chart of a method for determining a bridge path trajectory in one embodiment;

[0043] Figure 17 A schematic flow chart of a method for determining a bridge end point in one embodiment;

[0044] Figure 18 is a distribution structure diagram of multiple blood vessel segmentation points in a trajectory diagram in another embodiment;

[0045] Figure 19 A schematic flow chart of a method for determining a bridge end point in another embodiment;

[0046] Figure 20 A schematic flow chart of a specific method for obtaining bridge reconstruction information in another embodiment;

[0047] Figure 21 FIG. 4 is a structural block diagram of a bridge reconstruction device in one embodiment. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0049] The bridge reconstruction method provided in this application can be applied to Figure 1 The computer equipment shown. Figure 1 As shown, the computer device includes a processor, memory, network interface, display screen, and input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store cardiac images, vascular segmentation results, and pre-trained segmentation models. The network interface of the computer device is used to communicate with an external terminal via a network connection. When executed by the processor, the computer program implements a bypass reconstruction method.

[0050] Those skilled in the art will understand that Figure 1 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0051] It should be noted that the bridge reconstruction method provided in the embodiments of this application may be performed by a bridge reconstruction device, which may be implemented as part or all of a computer device through software, hardware, or a combination of software and hardware. The following method embodiments are described using a computer device as an example.

[0052] Among them, in possible implementations, the above-mentioned bypass reconstruction method can be implemented in various ways. Specifically, several bypass reconstruction methods can be further described in detail through the technical solutions in the following embodiments, but are not limited to the following embodiments.

[0053] The first bridge reconstruction method

[0054] The bridge reconstruction method can be described in detail through the following examples:

[0055] like Figure 2 The figure shows a flow chart of a bridge reconstruction method provided by an embodiment, wherein the method is applied to Figure 1 The computer device in the example is used to illustrate the process, including the following steps:

[0056] S1000: Obtain heart segmentation results and blood vessel segmentation results.

[0057] Specifically, the computer device can input multiple frames of cardiac images of the diagnosis and treatment object into a pre-trained cardiac segmentation model to obtain a cardiac segmentation result. Further, the vascular range is determined by the cardiac segmentation result, and the vascular range is segmented to obtain a vascular segmentation result. Optionally, the vascular segmentation result can be a segmentation result corresponding to all vascular regions in the cardiac segmentation result. The above-mentioned cardiac image can be a three-dimensional CT enhanced image in DICOM format. The above-mentioned cardiac segmentation result can be a cardiac segmentation image containing each chamber of the heart, the aorta, and the aortic arch, and each chamber of the heart, the aorta, and the aortic arch in the cardiac segmentation image have corresponding labels to distinguish different tissues in the cardiac segmentation image. The various chambers of the heart can be the left atrium, right atrium, left ventricle, and right ventricle of the heart, wherein the labels corresponding to the left atrium, right atrium, left ventricle, right ventricle, aorta, and aortic arch of the heart can be different. In addition, the above-mentioned cardiac segmentation result can also be represented by cardiac segmentation data containing each chamber of the heart, the aorta, and the aortic arch.

[0058] It can be understood that the above-mentioned pre-trained heart segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model; wherein, the computer device can perform network model training on the initial heart segmentation model through a heart image training set to obtain a pre-trained heart segmentation model.

[0059] At the same time, the heart segmentation result can be taken as a whole to determine the blood vessel range. The computer device can extend the heart segmentation result outward by a certain preset range to obtain the blood vessel range, and further input the blood vessel range into a pre-trained blood vessel segmentation model to obtain the blood vessel segmentation result. The above preset range can be customized according to actual conditions, as long as the blood vessel range is smaller than the range of the heart image. If the heart segmentation result is represented by a heart segmentation image, the blood vessel range can be an image representation obtained by extending the heart segmentation image outward by a certain preset range; if the heart segmentation result is represented by heart segmentation data, the blood vessel range can be a heart segmentation data representation corresponding to the image obtained by extending the heart segmentation image outward by a certain preset range. The above blood vessel segmentation result can be a binary blood vessel segmentation result.

[0060] The pre-trained vessel segmentation model may be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained vessel segmentation model and the pre-trained heart segmentation model have the same structure, their network parameters may differ after network model training. The computer device may perform network model training on the initial vessel segmentation model using a vessel range training set to obtain the pre-trained vessel segmentation model.

[0061] In addition, both the heart segmentation model and the blood vessel segmentation model can complete network model training before executing S100 in this embodiment.

[0062] S1100 , obtaining a target blood vessel mask based on the heart segmentation result, the blood vessel segmentation result, and the segmentation model.

[0063] Specifically, the computer device may process the heart segmentation result and the blood vessel segmentation result to obtain a processed result, and input the processed result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. Alternatively, the computer device may directly input the heart segmentation result and the blood vessel segmentation result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. In this embodiment, the heart segmentation result and the blood vessel segmentation result may both be understood as images or data.

[0064] It is understood that the segmentation model can be a pre-trained neural network model. The segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained segmentation model has the same structure as the pre-trained vascular segmentation model and the pre-trained heart segmentation model, the network parameters of the three models can be different after network model training. Furthermore, the segmentation model can complete network model training before executing S200 in this embodiment. Specifically, the computer device can input the heart segmentation results, the heart segmentation results from the vascular segmentation training set, and the vascular segmentation results into the initial segmentation model to obtain a vascular prediction mask. The computer device can calculate the prediction error between the vascular prediction mask and the standard vascular mask using a loss function, and update the initial network parameters of the initial segmentation model based on the prediction error. The above training steps are continuously iterated until the prediction error meets a preset error threshold or the number of iterations reaches a preset iteration threshold, thereby obtaining the pre-trained segmentation model. The standard vascular mask can be an idealized vascular mask corresponding to the entire vascular vessel, i.e., the gold standard for network training.

[0065] S1200: Obtain bypass reconstruction information based on the target blood vessel mask.

[0066] Specifically, the computer device can perform transformation, comparison, calculation, screening, and / or analysis on the target vessel mask to obtain bypass reconstruction information. Alternatively, the computer device can first filter target bypass data from the target vessel mask according to preset conditions and then analyze the target bypass data to determine the bypass reconstruction information. The target vessel mask can be understood as a multi-vessel label mask corresponding to the vessel segmentation result. The bypass reconstruction information can include at least one of the bypass trajectory, the range of the bypass trajectory, the bypass starting and ending points in the bypass trajectory, the bypass trajectory size, and each path point in the bypass trajectory.

[0067] S1300: Process the bridging reconstruction information to obtain a bridging reconstruction result.

[0068] Specifically, the computer device can process the bypass reconstruction information to obtain a three-dimensional heart bypass model, i.e., the bypass reconstruction result. Alternatively, the computer device can first segment normal blood vessels to obtain centerline information for all vessels, and then process the centerline information of all vessels and the bypass reconstruction information to obtain the bypass reconstruction result. Furthermore, the computer device can output the bypass reconstruction result and display it to the physician, allowing the physician to conduct regular follow-up visits with the patient based on the bypass reconstruction results.

[0069] In the above-mentioned bypass reconstruction method, the computer equipment can obtain the heart segmentation results and the blood vessel segmentation results, obtain the target blood vessel mask based on the heart segmentation results, the blood vessel segmentation results and the segmentation model, obtain the bypass reconstruction information based on the target blood vessel mask, and process the bypass reconstruction information to obtain the bypass reconstruction result; this method obtains the bypass reconstruction information through a neural network model, and determines the bypass reconstruction result through the reconstruction of the bypass reconstruction information, which can avoid manual reconstruction of the bypass model, save human resources and bypass reconstruction time, reduce bypass reconstruction costs, and further improve bypass reconstruction efficiency; at the same time, this method realizes bypass reconstruction through a neural network model, which can avoid reconstruction errors caused by manual participation in the reconstruction process, thereby improving the accuracy of bypass reconstruction.

[0070] As one example, Figure 3 As shown, the blood vessel segmentation result includes a first blood vessel segmentation result, and the segmentation model includes a first segmentation model. The step of obtaining a target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result, and the segmentation model in S1100 can be implemented by the following steps:

[0071] S1110 . Determine a distance field using the heart mask, where the distance field includes a distance field of the heart chamber and a distance field of the aorta and aortic arch.

[0072] Specifically, the heart mask may include a heart chamber mask and an aorta and aortic arch mask, and the heart mask, heart chamber mask, and aorta and aortic arch mask are all the same size as the heart segmentation image. The computer device may generate a heart chamber mask based on the labels corresponding to the left ventricle and right ventricle of the heart in the heart segmentation image. The pixel values ​​corresponding to the left ventricle and right ventricle regions of the heart in the heart chamber mask may be 1, while the pixel values ​​of the remaining regions may be 0. Furthermore, the computer device may also generate an aorta and aortic arch mask based on the labels corresponding to the aorta and aortic arch in the heart segmentation image. The pixel values ​​corresponding to the aorta and aortic arch regions of the heart in the generated aorta and aortic arch mask may be 1, while the pixel values ​​of the remaining regions may be 0.

[0073] It should be noted that the left and right ventricles of the heart in the cardiac chamber mask can be considered as a whole and referred to as the left and right ventricles; and the aorta and aortic arch in the aortic arch mask can be considered as a whole and referred to as the aorta. Furthermore, the computer device can use Euclidean distance, Manhattan distance, cosine distance, Minkowski distance, or Chebyshev distance calculation methods to calculate the distance between each point in the cardiac chamber mask, excluding the region where the left and right ventricles are located, and the nearest point on the surface of the left and right ventricles, and use all the calculated distances as the distance field of the cardiac chamber. Simultaneously, the computer device can also use Euclidean distance, Manhattan distance, cosine distance, Minkowski distance, or Chebyshev distance calculation methods to calculate the distance between each point in the aortic arch mask, excluding the region where the aorta is located, and the nearest point on the surface of the aorta, and use all the calculated distances as the distance field of the aorta.

[0074] S1120 , obtaining an initial blood vessel mask through the heart segmentation result, the distance field, the first blood vessel segmentation result, and the first segmentation model.

[0075] Specifically, the computer device can process the heart segmentation result and the first blood vessel segmentation result, and then input the processed result and the distance field into the first segmentation model for processing to obtain an initial blood vessel mask. Alternatively, the computer device can directly input the heart segmentation result, the distance field, and the first blood vessel segmentation result into the first segmentation model to obtain an initial blood vessel mask. The initial blood vessel mask is part of the target blood vessel mask, and the target blood vessel mask and the initial blood vessel mask may include a blood vessel mask and a bypass mask. The above-mentioned first blood vessel segmentation result can be a segmentation result corresponding to a partial blood vessel area in the heart segmentation result, or a segmentation result in which the blood vessel range is only in the vicinity of the heart's chamber; the distance field input into the first segmentation model can be the distance field of the heart chamber and / or the distance field of the aorta and aortic arch.

[0076] It should be noted that the computer device can perform network model training on the initial first segmentation model through the training set of the distance field of the heart chamber and the distance field of the aorta and the aortic arch, the heart segmentation result and the blood vessel segmentation result, to obtain a pre-trained first segmentation model. Specifically, the computer device can input the distance field of the heart chamber and the distance field of the aorta and the aortic arch in the training set, the heart segmentation result and the blood vessel segmentation result corresponding to the blood vessel area around the heart chamber in the training set into the initial first segmentation model. The initial first segmentation model First, the heart segmentation results and the vascular segmentation results are mapped and merged to obtain a merged segmentation result. Then, each point in the vascular area in the merged segmentation result is voted to obtain a multi-vessel prediction mask. The prediction error value between the multi-vessel prediction mask and the standard multi-vessel mask is calculated using a loss function, and the initial network parameters in the initial first segmentation model are updated based on the prediction error value. The above training steps are continuously iterated until the prediction error value meets the preset error threshold or the number of iterations reaches the preset iteration threshold, thereby obtaining the pre-trained first segmentation model. When the first segmentation model is used, the first segmentation model ultimately outputs the initial vascular mask.

[0077] In addition, when the network model is trained on the initial first segmentation model, the images in the training set of the heart segmentation results and the corresponding blood vessel segmentation results can be images with a size between 64*64*64 and 256*256*256. In this embodiment, the image size can be selected as 128*128*128, and the image resolution can be selected from 0.6mm to 2.0mm. In this embodiment, the image resolution can be 1.2mm. During the network model training process, the minimization of loss functions such as cross entropy, focal loss and / or diceloss can be used as the optimization goal, and parameter optimization methods such as Adam, sgd, AdamW, and RMSprop can be used to achieve optimization processing. The above-mentioned standard multi-vessel mask can be an idealized multi-vessel mask. The structure of the first segmentation model and the normal blood vessel segmentation model can be the same.

[0078] This embodiment uses the distance field of the heart chambers and the distance field of the aorta and aortic arch to help the first segmentation model determine the relative position of blood vessels within the heart or aorta, thereby assisting in improving blood vessel segmentation. The first segmentation model can include four input channels and 17 output channels. The four input channels are the heart segmentation result input channel, the first blood vessel segmentation result input channel, the heart chamber distance field input channel, and the aorta and aortic arch distance field input channel. The 17 output channels can output 15 types of blood vessel masks, a bypass mask, and a background mask corresponding to the non-vascular regions in the heart segmentation result.

[0079] S1130: Determine a target blood vessel mask based on the initial blood vessel mask.

[0080] Specifically, the computer device may perform expansion processing on the initial vascular mask to obtain the target vascular mask, or perform processing on the initial vascular mask, the heart segmentation result, the vascular segmentation result, and the segmentation model to obtain the target vascular mask.

[0081] The above-mentioned bypass reconstruction method can determine the distance field through the heart mask, which includes the distance field of the heart chamber and the distance field of the aorta and aortic arch. The initial blood vessel mask is obtained through the heart segmentation result, the distance field, the first blood vessel segmentation result and the first segmentation model. The target blood vessel mask is determined based on the initial blood vessel mask. This method obtains the target blood vessel mask of the entire blood vessel area through the neural network model, thereby improving the accuracy of the obtained blood vessel mask and reducing the time for determining the blood vessel mask.

[0082] As one example, Figure 4 As shown, the step of obtaining the initial blood vessel mask in the above S1120 using the heart segmentation result, the distance field, the first blood vessel segmentation result, and the first segmentation model includes:

[0083] S1121 , mapping and merging the dirty segmentation result and the first blood vessel segmentation result to obtain a merged segmentation result.

[0084] In this embodiment, the computer device may map each tissue in the heart segmentation result with the corresponding tissue in the first blood vessel segmentation result, and merge the heart segmentation result with the mapped corresponding first blood vessel segmentation result to obtain a merged segmentation result.

[0085] S1122: Input the distance field and the merged segmentation results into the first segmentation model to obtain an initial blood vessel mask.

[0086] Furthermore, the computer device may input the distance field and the merged segmentation result into the first segmentation model for processing to obtain an initial blood vessel mask, and the distance field may include the distance field of the heart chamber and the distance field of the aorta and aortic arch.

[0087] It is understood that the initial vessel mask or the target vessel mask may include at least one vessel mask and a bypass mask. The vessel mask may be any of 15 vessel masks, including the right coronary artery (RCA) mask, the right posterior descending coronary artery (R-PDA) mask, the left ventricular right posterior branch (R-PLB) mask, the left main coronary artery (LM) mask, the left anterior descending artery (LAD) mask, the diagonal branch (D) mask, the circumflex artery (LCX) mask, the obtuse marginal branch (OM) mask, the left posterior descending artery (L-PDA) mask, the left posterior left ventricular branch (L-PLB) mask, the intermediate branch (RAMUS) mask, the atrial branch (LACX) mask, the acute marginal branch (AM) mask, the anterior septal branch (S) mask, or the sinoatrial node branch (LSN) mask.

[0088] The above-mentioned bypass reconstruction method can map and merge the heart segmentation result and the first blood vessel segmentation result to obtain a merged segmentation result, and input the distance field and the merged segmentation result into the first segmentation model to obtain an initial blood vessel mask, and then determine the target blood vessel mask based on the initial blood vessel mask; this method obtains the blood vessel mask of the blood vessel area through a neural network model, thereby improving the accuracy of the obtained blood vessel mask and reducing the time for determining the blood vessel mask.

[0089] As one embodiment, the blood vessel segmentation result includes a second blood vessel segmentation result, and the segmentation model includes a second segmentation model; as shown in FIG. Figure 5 As shown, the step of determining the target blood vessel mask based on the initial blood vessel mask in S1130 can be implemented by the following steps:

[0090] S1131 : Determine whether the first blood vessel segmentation result includes bridging data based on the initial blood vessel mask.

[0091] Specifically, the computer device can determine all first vessel segmentation results belonging to the bypass mask in the merged segmentation result based on the initial vessel mask corresponding to the vessel region in the merged segmentation result, and then determine whether the first vessel segmentation result contains bypass data, i.e., bypass data, based on the number of all first vessel segmentation results belonging to the bypass mask. The number of all first vessel segmentation results belonging to the bypass mask can be equal to all first vessel segmentation results in the merged segmentation result, or equal to some first vessel segmentation results in the merged segmentation result, or there can be no first vessel segmentation results belonging to the bypass mask in the merged segmentation result. In other words, the number of all first vessel segmentation results belonging to the bypass mask can be equal to 0, or greater than or equal to 1.

[0092] S1132: If the first blood vessel segmentation result includes bypass data, obtain the remaining blood vessel segmentation results within the blood vessel range to obtain a second blood vessel segmentation result.

[0093] Specifically, when the computer device determines that the first vessel segmentation result includes bypass data, it can obtain the remaining vessel segmentation results within the vessel range to obtain a second vessel segmentation result. The second vessel segmentation result corresponds to the vessel region included in the area above the heart in the heart segmentation result. In other words, the result of combining the first and second vessel segmentation results can be equivalent to the image or data of the entire vessel region in the heart segmentation result.

[0094] S1133 , input the heart segmentation result, the distance field, the first blood vessel segmentation result, and the second blood vessel segmentation result into the second segmentation model to obtain a target blood vessel mask.

[0095] Specifically, the computer device can perform computational processing on the heart segmentation result, the first vessel segmentation result, and the second vessel segmentation result, and then input the computational processing result and the distance field into the second segmentation model for processing to obtain a target vessel mask. Alternatively, the computer device can directly input the distance field, the heart segmentation result, the first vessel segmentation result, and the second vessel segmentation result into the second segmentation model. The second segmentation model first maps and merges each tissue in the heart segmentation result with the corresponding tissue in the first and second vessel segmentation results to obtain a merged segmentation result. The second segmentation model then processes the distance field and the merged segmentation result to obtain a target vessel mask. The merged segmentation result obtained by mapping and merging the first and second vessel segmentation results can be a segmentation result corresponding to all vascular regions in the heart segmentation result. The distance field input into the second segmentation model can be a distance field of a heart chamber and / or a distance field of the aorta and aortic arch.

[0096] It should be noted that the computer device can perform network model training on the initial second segmentation model through the training set of the distance field of the heart chamber and the distance field of the aorta and the aortic arch, the heart segmentation result and the training set of the whole blood vessel segmentation result to obtain a pre-trained second segmentation model. Specifically, the computer device can input the distance field of the heart chamber and the distance field of the aorta and the aortic arch in the training set, the heart segmentation result and the blood vessel segmentation result of the whole blood vessel area in the training set into the initial second segmentation model. The initial second segmentation model first trains the heart segmentation result and the whole blood vessel segmentation result in the training set. The segmentation result is mapped and merged with the vessel segmentation result of the entire vessel region to obtain a merged segmentation result. A vote is then performed on each point in the vessel region in the merged segmentation result to obtain a multi-vessel prediction mask. A prediction error between the multi-vessel prediction mask and a standard multi-vessel mask is calculated using a loss function. Initial network parameters in the initial second segmentation model are updated based on the prediction error. The above training steps are iterated until the prediction error meets a preset error threshold or the number of iterations reaches a preset iteration threshold, thereby obtaining a pre-trained second segmentation model. When the second segmentation model is used, the second segmentation model ultimately outputs the target vessel mask. In this embodiment, all vessel segmentation results correspond to the vessel region corresponding to the combination of the first vessel segmentation result and the second vessel segmentation result.

[0097] It should be noted that the second segmentation model can include five input channels and 17 output channels. The five input channels are respectively an input channel for the heart segmentation result, an input channel for the first blood vessel segmentation result, an input channel for the second blood vessel segmentation result, an input channel for the distance field of the heart chambers, and an input channel for the distance field of the aorta and aortic arch. The 17 output channels can respectively output the 15 masks described above, the bypass mask, and the background mask corresponding to the non-vascular region in the heart segmentation result. In this embodiment, the structures of the first and second segmentation models can be identical, as long as the sampling ratio of the bypass data is increased during the training of the second segmentation model, so that the second segmentation model has better segmentation performance for the bypass data. However, the network parameters of the first and second segmentation models are different.

[0098] The above-mentioned bypass reconstruction method can, upon determining that the first blood vessel segmentation result contains bypass data, continue to obtain a second blood vessel segmentation result by obtaining the remaining blood vessel segmentation results within the blood vessel range, and combine the first blood vessel segmentation result and the second blood vessel segmentation result to obtain a blood vessel segmentation result corresponding to the entire blood vessel area. The heart segmentation result, the distance field, and the blood vessel segmentation result corresponding to the entire blood vessel area are further processed using a neural network model to obtain a target blood vessel mask for the entire blood vessel area, thereby improving the integrity and accuracy of the obtained blood vessel mask. The neural network model processing can also shorten the time for determining the blood vessel mask, further improving the efficiency of bypass reconstruction.

[0099] As one example, Figure 6 As shown, the step of determining whether the first blood vessel segmentation result includes bridging data based on the initial blood vessel mask in the above S1131 can be implemented by the following steps:

[0100] S1131a. Determine a vascular connectivity domain based on the first vascular segmentation result, and obtain relevant information about the vascular connectivity domain; the relevant information about the vascular connectivity domain includes an initial spacing distance of the vascular connectivity domain in the distance field of the heart chamber or a size of the vascular connectivity domain.

[0101] Specifically, the computer device may select any one of all the vessel segmentation points corresponding to all the first vessel segmentation results in the merged segmentation result as a reference point, obtain a preset number of vessel segmentation points surrounding the reference point, and determine the reference point and all the vessel segmentation points that form the vessel mask among the preset number of vessel segmentation points surrounding the reference point based on the initial vessel mask. The reference point and all the vessel segmentation points that form the vessel mask among the preset number of vessel segmentation points are then defined as a vessel connected domain. Simultaneously, the computer device may traverse all the vessel segmentation points corresponding to all the first vessel segmentation results in the merged segmentation result and determine the vessel connected domain corresponding to each vessel segmentation point. If the first vessel segmentation result corresponds to a three-dimensional image, the vessel connected domain may be determined by selecting the preset number of vessel segmentation points plus one vessel segmentation point from all the vessel segmentation results corresponding to the first vessel segmentation results (i.e., the reference point and the preset number of vessel segmentation points surrounding the reference point). Within the vessel connected domain to which any vessel segmentation point belongs, at least one other vessel segmentation point can be found to be connected to the current vessel segmentation point, i.e., at least one other vessel segmentation point can be found to be connected to the current vessel segmentation point within the same neighborhood (the preset number of surrounding vessel segmentation points, i.e., the preset number of neighborhoods). The above-mentioned preset number can be any value, as long as it ensures that any blood vessel segmentation point can find at least one other blood vessel segmentation point to connect with it within the blood vessel connection domain to which it belongs.

[0102] The above-mentioned blood vessel segmentation result may be blood vessel segmentation data. The blood vessel segmentation data may be in the form of point coordinates, which are displayed in a coordinate system, and each blood vessel segmentation result has a corresponding blood vessel segmentation point.

[0103] It will be appreciated that the computer device can determine the distance corresponding to each vascular connectivity domain in the distance field of the cardiac chamber, i.e., the initial separation distance, and can also obtain the size of the vascular connectivity domain. The size of the vascular connectivity domain can be equal to the total number of vascular segmentation points contained in the vascular connectivity domain. In this embodiment, the number of initial separation distances corresponding to each vascular connectivity domain can be equal to the number of distances contained in the distance field of the cardiac chamber.

[0104] S1131b: Determine whether the first blood vessel segmentation result includes bridging data based on the initial blood vessel mask and related information of the blood vessel connected domain.

[0105] Specifically, the computer device can perform comparison processing, calculation processing, screening processing and / or analysis processing based on the relevant information of the initial blood vessel mask and the blood vessel connected domain to determine whether all the first blood vessel segmentation results in each blood vessel connected domain contain bridging data.

[0106] The above-mentioned bypass reconstruction method can determine the vascular connectivity domain based on the first vascular segmentation result and obtain relevant information of the vascular connectivity domain. Based on the initial vascular mask and the relevant information of the vascular connectivity domain, it is determined whether the first vascular segmentation result contains bypass data. When it is determined that the first vascular segmentation result in the vascular connectivity domain contains bypass data, the target vascular mask corresponding to the entire blood vessel in the heart segmentation result is obtained, thereby improving the accuracy and completeness of the obtained vascular mask.

[0107] As one embodiment, the initial vessel mask includes a bridging mask; Figure 7 As shown, the step of determining whether the first blood vessel segmentation result includes bridging data based on the initial blood vessel mask and the related information of the blood vessel connected area in the above S1131b may specifically include:

[0108] S1131c: If the initial interval distance is greater than a first preset distance threshold or the size of the vascular connected domain is greater than a first preset number threshold, obtain a first total number of first vascular segmentation results corresponding to the bridging masks contained in other vascular connected domains.

[0109] Specifically, for each vascular connectivity domain, the computer device can determine whether the initial interval distances corresponding to the current vascular connectivity domain in the distance field of the cardiac chamber are all greater than a first preset distance threshold. If the judgment result is yes, the current vascular connectivity domain is not considered. At this time, other vascular connectivity domains outside the current vascular connectivity domain can be obtained, and a first total number of vascular segmentation results corresponding to all bridging masks contained in each of the other vascular connectivity domains can be obtained; each of the other vascular connectivity domains has a corresponding first total number, and the first total number can be greater than or equal to 1.

[0110] Meanwhile, the first total number can also be determined in another manner. Specifically, the computer device can first calculate the size of each vascular connectivity domain and determine whether the size of the current vascular connectivity domain is greater than a first preset number threshold. If the determination result is yes, the current vascular connectivity domain is disregarded, and the first total number of vascular segmentation results corresponding to all bridging masks contained in each vascular connectivity domain other than the current vascular connectivity domain is obtained. The size of the vascular connectivity domain can be understood as the total number of vascular segmentation points contained in the vascular connectivity domain.

[0111] In addition, for each vascular connectivity domain, the computer device can also determine whether the initial interval distances corresponding to the current vascular connectivity domain in the distance field of the cardiac chamber are greater than or equal to a first preset distance threshold. If the judgment result is yes, the current vascular connectivity domain is not considered. At this time, the other vascular connectivity domains outside the current vascular connectivity domain can be obtained, and the first total number of vascular segmentation results corresponding to all bypass masks contained in each other vascular connectivity domain is obtained; or, the size of each vascular connectivity domain can be calculated first, and it can be determined whether the size of the current vascular connectivity domain is greater than or equal to the first preset number threshold. If the judgment result is yes, the current vascular connectivity domain is not considered, and the first total number of vascular segmentation results corresponding to all bypass masks contained in each other vascular connectivity domain outside the current vascular connectivity domain is obtained.

[0112] It is understandable that the specific values ​​of the above-mentioned first preset distance threshold and the first preset quantity threshold are not limited, and can be customized by the user according to actual needs, or can be obtained by cross-validation and taking the average value of each verification result. Among them, for different types of preset thresholds, cross-validation can be understood as dividing the data set corresponding to the preset threshold into multiple parts, taking one of them as the verification set, and the remaining data sets as the training set. Each time the corresponding threshold is adjusted on the training set to achieve the optimal value, and the threshold corresponding to the maximum accuracy obtained in the verification is used as the optimal result of each verification, that is, the preset threshold. For example, when determining the first preset distance threshold, the data set can be a set corresponding to multiple initial interval distances.

[0113] S1131d. If the first total number is greater than a second preset number threshold, determine the first blood vessel segmentation result corresponding to the bridging mask contained in the other blood vessel connected domain as bridging data, and determine that the first blood vessel segmentation result in the other blood vessel connected domain includes bridging data.

[0114] Furthermore, the computer device may continue to determine whether the first total number of first vessel segmentation results corresponding to the bypass masks contained in other connected vascular domains is greater than a second preset number threshold. If so, the first vessel segmentation results corresponding to all bypass masks contained in the other connected vascular domains may be determined as bypass data. In this case, the first vessel segmentation results in the other connected vascular domains are directly determined to contain bypass data. Alternatively, the computer device may also determine whether the first total number of first vessel segmentation results corresponding to the bypass masks contained in other connected vascular domains is equal to a second preset number threshold. If so, the first vessel segmentation results corresponding to all bypass masks contained in the other connected vascular domains may be determined as bypass data. In this case, the first vessel segmentation results in the other connected vascular domains are directly determined to contain bypass data. Optionally, the second preset number threshold and the first preset number threshold may be equal, and the specific determination method may be the same. In this embodiment, steps S1131b and S1131d are performed for each connected vascular domain, that is, determining whether the first vessel segmentation result in each connected vascular domain contains bypass data.

[0115] The above-mentioned bypass reconstruction method can determine whether the first blood vessel segmentation result contains bypass data, and then when it is determined that the first blood vessel segmentation result in the blood vessel connectivity domain contains bypass data, obtain the target blood vessel mask corresponding to the entire blood vessel in the heart segmentation result, thereby improving the accuracy and completeness of the obtained blood vessel mask.

[0116] As one of the embodiments, the step of obtaining the bypass reconstruction information based on the target blood vessel mask in the above S1200 may include: performing post-processing on the target blood vessel mask and the blood vessel segmentation result to obtain the bypass starting point, bypass path and anastomosis between the bypass and the normal blood vessel in the bypass reconstruction information.

[0117] Specifically, the post-processing may be at least one of comparison processing, calculation processing, screening processing, analysis processing, and the like. That is, the computer device may perform transformation processing, comparison processing, calculation processing, screening processing, and / or analysis processing on the target vessel mask, the first vessel segmentation result, and the second vessel segmentation result to obtain the bypass reconstruction information, or may merge the first vessel segmentation result and the second vessel segmentation result based on the target vessel mask, and then determine the bypass reconstruction information based on the merged result.

[0118] To reduce computational complexity, bypass reconstruction can be achieved by simply determining the bypass starting point, bypass trajectory, and anastomosis between the bypass and normal blood vessels in the bypass reconstruction information. Therefore, in this embodiment, the bypass reconstruction information may include the bypass starting point, bypass trajectory, and anastomosis between the bypass and normal blood vessels. The number of bypass starting points, bypass trajectory, and anastomosis between the bypass and normal blood vessels can all be greater than one.

[0119] For example, Figure 7a The diagram shows multiple frames of cardiac images of a patient, cardiac segmentation results corresponding to the cardiac images, a first vessel segmentation result, an initial vessel mask corresponding to the first vessel segmentation result, the first and second vessel segmentation results (i.e., vessel segmentation results for the entire cardiac vessel region), target vessel masks corresponding to the first and second vessel segmentation results, and a corresponding bypass reconstruction result. Different vessel segmentation points correspond to different vessel masks, and therefore, the vessel segmentation points of different vessel masks have different label values. In the images corresponding to the initial and target vessel masks, the vessel segmentation points of different vessel masks will display different label values.

[0120] The above-mentioned bypass reconstruction method can obtain bypass reconstruction information by post-processing the complete blood vessel mask and the complete blood vessel segmentation result, thereby improving the completeness and accuracy of the bypass reconstruction information and further improving the accuracy of the bypass reconstruction result.

[0121] The second bridge reconstruction method

[0122] The bridge reconstruction method can be described in detail through the following examples:

[0123] like Figure 8 The figure shows a flow chart of a bridge reconstruction method provided by an embodiment, wherein the method is applied to Figure 1 The computer device in the example is used to illustrate the process, including the following steps:

[0124] S2000: Obtain heart segmentation results and blood vessel segmentation results.

[0125] Specifically, the computer device can input multiple frames of cardiac images of the diagnosis and treatment object into a pre-trained cardiac segmentation model to obtain a cardiac segmentation result. Further, the vascular range is drawn through the cardiac segmentation result, and the data corresponding to the vascular range is subjected to vascular segmentation to obtain a vascular segmentation result. Optionally, the vascular segmentation result can be a segmentation result corresponding to the entire vascular area in the cardiac segmentation result. The above-mentioned cardiac image can be a three-dimensional CT enhanced image in DICOM format. The above-mentioned cardiac segmentation result can be a cardiac segmentation image containing the various chambers of the heart, the aorta, and the aortic arch, and the various chambers of the heart, the aorta, and the aortic arch in the cardiac segmentation image have corresponding labels to distinguish different tissues in the cardiac segmentation image. The various chambers of the heart can be the left atrium, right atrium, left ventricle, and right ventricle of the heart. In addition, the above-mentioned cardiac segmentation result can also be represented by cardiac segmentation data containing the various chambers of the heart, the aorta, and the aortic arch.

[0126] It can be understood that the above-mentioned pre-trained heart segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model; wherein, the computer device can perform network model training on the initial heart segmentation model through a heart image training set to obtain a pre-trained heart segmentation model.

[0127] At the same time, the heart segmentation result can be taken as a whole to determine the blood vessel range. The computer device can extend the heart area range in the heart segmentation result outward by a certain preset range to obtain the blood vessel range, and further input the blood vessel range into a pre-trained blood vessel segmentation model to obtain the blood vessel segmentation result. The above preset range can be customized according to actual conditions, as long as the blood vessel range is smaller than the range of the heart image. If the heart segmentation result is represented by a heart segmentation image, the blood vessel range can be an image representation obtained by extending the heart segmentation image outward by a certain preset range; if the heart segmentation result is represented by heart segmentation data, the blood vessel range can be a heart segmentation data representation corresponding to the image obtained by extending the heart segmentation image outward by a certain preset range. The above blood vessel segmentation result can be a binary blood vessel segmentation result.

[0128] The pre-trained vessel segmentation model may be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained vessel segmentation model and the pre-trained heart segmentation model have the same structure, their network parameters may differ after network model training. The computer device may perform network model training on the initial vessel segmentation model using a vessel range training set to obtain the pre-trained vessel segmentation model.

[0129] In addition, both the heart segmentation model and the blood vessel segmentation model can complete network model training before executing S2000 in this embodiment.

[0130] S2100 , obtaining a target blood vessel mask based on the heart segmentation result, the blood vessel segmentation result, and the segmentation model.

[0131] Specifically, the computer device may process the heart segmentation result and the blood vessel segmentation result to obtain a processed result, and input the processed result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. Alternatively, the computer device may directly input the heart segmentation result and the blood vessel segmentation result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. In this embodiment, the heart segmentation result and the blood vessel segmentation result may both be understood as images or data.

[0132] It is understood that the segmentation model can be a pre-trained neural network model. The segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained segmentation model has the same structure as the pre-trained vascular segmentation model and the pre-trained heart segmentation model, the network parameters of the three models can be different after network model training. Furthermore, the segmentation model can complete network model training before executing S2100 in this embodiment.

[0133] It should be noted that the computer device can input the heart segmentation results and the data in the training set corresponding to the vascular segmentation results of the entire blood vessel within the heart segmentation results into the initial segmentation model to obtain a vascular prediction mask. The prediction error between the vascular prediction mask and the standard vascular mask is calculated using a loss function, and the initial network parameters in the initial segmentation model are updated based on the prediction error. The above training steps are continuously iterated until the prediction error meets a preset error threshold or the number of iterations reaches a preset iteration threshold, thereby obtaining a pre-trained segmentation model. The above-mentioned standard vascular mask can be an idealized vascular mask corresponding to the entire blood vessel. In this embodiment, the vascular prediction mask output by the segmentation model can be a vascular mask corresponding to the entire blood vessel.

[0134] S2200 : Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result. The bypass reconstruction information includes a bypass starting point and an anastomosis between the bypass and the normal blood vessel.

[0135] Specifically, the computer device can perform transformation processing, comparison processing, calculation processing, screening processing and / or analysis processing on the target vessel mask and the vessel segmentation result to obtain the bypass reconstruction information, or filter the target bypass data from the target vessel mask according to preset conditions, and analyze the target bypass data and the vessel segmentation result to determine the bypass reconstruction information. The above-mentioned target vessel mask can be understood as a multi-vessel label mask. The above-mentioned bypass reconstruction information can include at least one of the bypass path, the range of the bypass path, the bypass starting point and the bypass ending point in the bypass path, the size of the bypass path, and each path point in the bypass path. However, in this embodiment, the bypass reconstruction information can include the bypass starting point and the anastomosis between the bypass and the normal vessel. The above-mentioned anastomosis between the bypass and the normal vessel can be understood as the connection point between the bypass and the normal vessel.

[0136] S2300: Process the bridging reconstruction information to obtain a bridging reconstruction result.

[0137] Specifically, the computer device can process the bypass reconstruction information to obtain a three-dimensional heart bypass model, i.e., the bypass reconstruction result. Alternatively, the computer device can first segment normal blood vessels to obtain centerline information for all vessels, and then process the centerline information of all vessels and the bypass reconstruction information to obtain the bypass reconstruction result. Furthermore, the computer device can output the bypass reconstruction result and display it to the physician, allowing the physician to conduct regular follow-up visits with the patient based on the bypass reconstruction results.

[0138] In the above-mentioned bypass reconstruction method, the computer equipment can obtain the heart segmentation results and the blood vessel segmentation results, obtain the target blood vessel mask based on the heart segmentation results, the blood vessel segmentation results and the segmentation model, obtain the bypass reconstruction information based on the target blood vessel mask, and process the bypass reconstruction information to obtain the bypass reconstruction result; this method obtains the bypass reconstruction information through a neural network model, and determines the bypass reconstruction result through the reconstruction of the bypass reconstruction information, which can avoid manual reconstruction of the bypass model, save human resources and bypass reconstruction time, reduce bypass reconstruction costs, and further improve bypass reconstruction efficiency; at the same time, this method realizes bypass reconstruction through a neural network model, which can avoid reconstruction errors caused by manual participation in the reconstruction process, thereby improving the accuracy of bypass reconstruction.

[0139] As one example, Figure 9 As shown, the step of obtaining the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result in S2200 can be implemented by the following steps:

[0140] S2210. Obtain the blood vessel centerline based on the blood vessel segmentation result.

[0141] Specifically, the computer device can perform skeletonization processing on the vessel segmentation results to obtain at least one vessel centerline. All vessel segmentation results can be represented by a vessel segmentation image, and each vessel segmentation result can be understood as vessel segmentation data. The skeletonization process can be understood as reducing a binary object to a representation of 1 pixel width.

[0142] Alternatively, the computer device may determine the starting vessel layer of the starting vessel segment and the ending vessel layer of the ending vessel segment from the vessel segmentation image, and use the center of the starting vessel layer of the starting vessel segment as the starting point of the vessel centerline, and the center of the ending vessel layer of the ending vessel segment as the ending point of the vessel centerline. The vessel centerline is then extracted from the vessel segmentation image using the maximum inscribed sphere method based on the starting and ending points of the vessel centerline. Of course, other methods may also be used to obtain the vessel centerline from the vessel segmentation results, and this is not limited to these methods.

[0143] S2220. Determine candidate bridging points based on all vascular points on the vascular centerline.

[0144] Specifically, the computer device may obtain all vessel points on the centerline of each vessel and use some of the vessel points as candidate bridge points. The candidate bridge points may be some of the starting and ending points and / or end points on the centerline of each vessel.

[0145] Among them, such as Figure 10 As shown, the step of determining candidate bridging points based on all vascular points on the vascular centerline in S2220 may specifically include:

[0146] S2221. Obtain a blood vessel point set corresponding to all blood vessel points on the blood vessel centerline, and obtain blood vessel endpoints in the blood vessel point set.

[0147] It is understood that the computer device can obtain all vascular points on the vascular centerline and store these vascular points in a vascular point set. When storing, the vascular points can be stored together with their respective coordinates. At the same time, the computer device can also filter out the two endpoints of each vascular centerline from all vascular points and obtain the coordinates of the two endpoints on each vascular centerline. Furthermore, the computer device can sequentially traverse the endpoints of each vascular centerline and use each traversed endpoint as the vascular endpoint.

[0148] Optionally, each vascular centerline may correspond to a vascular point set, and the order of different vascular points stored in the vascular point set may be equal to the position order of all vascular points on the corresponding vascular centerline from any end point of the vessel to the other end point of the vessel.

[0149] S2222: Determine a preset number of target blood vessel points in the blood vessel point set that include blood vessel endpoints.

[0150] It should be noted that the computer device may determine, based on all vascular points in the vascular point set, a preset number of target vascular points in each vascular point set that are located before or after the vascular endpoint of the corresponding vascular centerline. This preset number may be less than or equal to the total number of all vascular points on the vascular centerline where the current vascular endpoint is located. All vascular points on each vascular centerline may include the two vascular endpoints on that vascular centerline.

[0151] S2223. Determine candidate bypass points based on the target blood vessel point.

[0152] Specifically, the computer device may determine whether a vessel endpoint among a preset number of target vessel points on each vessel centerline is a candidate bridging point based on a target vessel mask for the preset number of target vessel points. At least one candidate bridging point may exist among all vessel endpoints on all vessel centerlines. The candidate bridging point may be represented by vessel point coordinates.

[0153] This embodiment can screen out candidate bridging points from all vascular points on the vascular centerline, and further determine the bridging starting point based on the candidate bridging points. This process can narrow the candidate range of the bridging starting point, thereby reducing the amount of calculation to determine the bridging starting point and improving the speed of determining the bridging starting point.

[0154] S2230: Determine the starting point of the bridge based on the candidate bridge points.

[0155] Specifically, the computer device may directly determine the bridging candidate point as the bridging starting point, or determine some of all the bridging candidate points as the bridging starting point. In this embodiment, the steps S2221-S2224 above are performed for each blood vessel centerline.

[0156] The above-mentioned bridge reconstruction method can determine the bridge candidate points in the blood vessel segmentation results, and then determine the bridge starting point through the bridge candidate points. This method avoids manual participation, can save human resources and time for determining the bridge starting point, and reduce the cost of determining the bridge starting point. At the same time, no manual participation is required to determine the bridge starting point, which can improve the accuracy of the determined bridge starting point.

[0157] As one embodiment, the step of determining bypass candidate points based on the target vessel points in S2223 may include at least one of the following conditions: if a second total number of bypass vessel points among all target vessel points is greater than a third preset number threshold, determining the vessel endpoints among all target vessel points as bypass candidate points; if the distance between each target vessel point and the highest point in the heart chamber is greater than a second preset distance threshold, determining the vessel endpoints among all target vessel points as bypass candidate points; and if the minimum distance between the vessel endpoints among all target vessel points and the aortic arch is less than a third preset distance threshold, determining the vessel endpoint as a bypass candidate point.

[0158] Specifically, the computer device can obtain the total number of target vessel points belonging to the bypass mask from the preset number of target vessel points (i.e., the second total number), and at the same time, calculate the distance between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber, and then determine whether the second total number is greater than a third preset number threshold, and / or whether the distance between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber is greater than the second preset distance threshold. If the judgment result is yes, the vessel endpoints in the preset number of target vessel points can be determined as bypass candidate points; otherwise, if it is determined that the second total number is less than or equal to the third preset number threshold, and / or at least one of the distances between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber is less than or equal to the second preset distance threshold, the current preset number of target vessel points can be filtered out. That is, in this case, there are no bypass candidate points among the current preset number of target vessel points.

[0159] Of course, the computer device may also determine whether the second total number is equal to a third preset number threshold, and / or whether the distance between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber is greater than or equal to the second preset distance threshold. If the determination result is yes, the vessel endpoints in the preset number of target vessel points may be determined as candidate bypass points. Otherwise, if it is determined that the second total number is less than the third preset number threshold, and / or at least one of the distances between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber is less than the second preset distance threshold, the current preset number of target vessel points may be filtered out. In other words, in this case, no candidate bypass points exist in the current preset number of target vessel points. Optionally, the distance between each target vessel point in the preset number of target vessel points and the highest point in the heart chamber region may be understood as the distance above the heart, wherein the highest point in the heart chamber region may be the point corresponding to the maximum coordinate of the heart chamber in the coronal plane.

[0160] At the same time, the computer device can also calculate the minimum distance between the blood vessel endpoint and the aorta and aortic arch among a preset number of target blood vessel points, and whether the minimum distance between the blood vessel endpoint and the aorta and aortic arch is less than or equal to a third preset distance threshold. If the judgment result is yes, the blood vessel endpoint among the preset number of target blood vessel points can be determined as a candidate point for bypass surgery; otherwise, if the minimum distance between the blood vessel endpoint and the aorta and aortic arch is less than or equal to the third preset distance threshold, the current preset number of target blood vessel points can be filtered out, that is, in this case, there are no candidate points for bypass surgery among these current preset number of target blood vessel points.

[0161] It is understood that the third preset number threshold, the second preset distance threshold, and the third preset distance threshold may be equal or unequal, and may be determined based on actual circumstances. In this embodiment, the third preset number threshold may be greater than 1 / 2 of the preset number, which may be equal to 50, the second preset distance threshold may be equal to 0.5 cm, and the third preset distance threshold may be equal to 1.5 cm.

[0162] The above-mentioned bridge reconstruction method can determine the candidate bridge points in the blood vessel segmentation results, thereby narrowing the candidate range of the bridge starting point, reducing the amount of calculation to determine the bridge starting point, and improving the speed of determining the bridge starting point, so that the bridge starting point can be quickly obtained in a short time.

[0163] As one example, Figure 11 As shown, the above-mentioned S2230 determines the bridge starting point based on the bridge candidate point, which can be achieved by the following steps:

[0164] S2231. Determine the target blood vessel connectivity domain based on the blood vessel segmentation result.

[0165] Specifically, the computer device may select any one of all the vessel segmentation points corresponding to the vessel segmentation result of the entire vessel region as a reference point, obtain a preset number of vessel segmentation points surrounding the reference point, and determine the reference point and all the vessel segmentation points that form the vessel mask among the preset number of vessel segmentation points surrounding the reference point based on the target vessel mask. The reference point and all the vessel segmentation points that form the vessel mask among the preset number of vessel segmentation points are then defined as a target vessel connected domain. Simultaneously, the computer device may traverse each vessel segmentation point corresponding to the vessel segmentation result and determine the target vessel connected domain corresponding to each vessel segmentation point. If the vessel segmentation result corresponds to a three-dimensional image, the target vessel connected domain may be determined by selecting a preset number of vessel segmentation points plus one from all the vessel segmentation points corresponding to the vessel segmentation result (i.e., the reference point and the preset number of vessel segmentation points surrounding the reference point). Within the target vessel connected domain to which any vessel segmentation point belongs, at least one other vessel segmentation point can be found that is connected to the selected current vessel segmentation point. That is, at least one other vessel segmentation point can be found that is connected to the current vessel segmentation point within the same neighborhood (the preset number of surrounding vessel segmentation points, i.e., the preset number of neighborhoods). In this embodiment, the preset number can be any value, as long as it ensures that any blood vessel segmentation point can find at least one other blood vessel segmentation point to connect with it within the target blood vessel connection domain.

[0166] It should be noted that each target vascular connectivity domain may be a collection of multiple vascular segmentation points.

[0167] S2232: Determine the target vessel connected domain to which the candidate bridging point belongs based on the candidate bridging point and the target vessel connected domain.

[0168] Specifically, for each candidate bridging point, the computer device may match the candidate bridging point with a vessel segmentation point within the target vessel connectivity domain, and determine the target vessel connectivity domain to which the successfully matched candidate bridging point belongs. Optionally, the target vessel connectivity domain to which the candidate bridging point belongs may be any one of all target vessel connectivity domains.

[0169] S2233. Obtain the distance between any two candidate bridging points within the target blood vessel connectivity domain.

[0170] It is understood that the computer device can use Euclidean distance, Manhattan distance, cosine distance, Minkowski distance, or Chebyshev distance calculation methods to calculate the interval distance between each pair of bridging candidate points among all bridging candidate points within the target vessel connectivity domain to which the bridging candidate point belongs. The target vessel connectivity domain to which the bridging candidate point belongs may include at least two bridging candidate points. If the target vessel connectivity domain to which the bridging candidate point belongs includes n (n>2) bridging candidate points, a distance can be determined for every two bridging candidate points among the n bridging candidate points, and n*(n-1) / 2 distances can be determined for the n bridging candidate points.

[0171] S2234. Determine the starting point of the bridge based on the interval distance.

[0172] It should be noted that the computer device may compare the separation distance with a preset distance threshold and determine whether the candidate bridge point is the bridge starting point based on the comparison result. The comparison result may be that the separation distance is greater than, less than, or equal to the preset distance threshold. In this embodiment, the preset distance threshold can be any value and is not limited to this.

[0173] The above-mentioned bridge reconstruction method can determine the starting point of the bridge through the candidate bridge point, and then obtain the bridge reconstruction result through processing the starting point of the bridge; this method avoids manual participation, can save human resources and time for determining the starting point of the bridge, and reduce the cost of determining the starting point of the bridge. At the same time, no manual participation is required to determine the starting point of the bridge, which can improve the accuracy of the determined starting point of the bridge, and further improve the accuracy of the bridge reconstruction result; in addition, this method can determine the starting point of the bridge within a smaller range, thereby reducing the amount of calculation for determining the starting point of the bridge, improving the speed of determining the starting point of the bridge, and can quickly obtain the starting point of the bridge in a short time, thereby improving the efficiency of bridge reconstruction.

[0174] As one example, Figure 12 As shown, the step of determining the starting point of the bridge according to the interval distance in the above S2234 may specifically include the following steps:

[0175] S2234a: If the interval distance is less than a fourth preset distance threshold, retain the bypass candidate point closest to the aorta among the bypass candidate points corresponding to the interval distance to obtain a first retained candidate point.

[0176] It should be noted that the computer device can determine whether the interval distance corresponding to each bypass candidate point is less than or equal to the fourth preset distance threshold. If the judgment result is yes, the bypass candidate point closest to the aorta among all the bypass candidate points corresponding to the interval distance is retained, and the retained bypass candidate point is used as the first retained candidate point. The interval distance corresponding to the above-mentioned bypass candidate point can be the interval distance between any two bypass candidate points in all the bypass candidate points in the target vascular connectivity domain to which the bypass candidate point belongs. The above-mentioned fourth preset distance threshold, the second preset distance threshold and the third preset distance threshold can be equal or unequal, and can all be determined based on actual conditions. In this embodiment, the above-mentioned fourth preset distance threshold can be equal to 0.5 cm.

[0177] S2234b. If the distance between the bypass candidate point within the target vascular connectivity domain and the outer surface of the heart chamber is greater than a fifth preset distance threshold, then, based on the target vascular mask, obtain a third total number of bypass candidate points belonging to the left anterior descending branch mask and / or the diagonal branch mask among all vascular segmentation points within the target vascular connectivity domain; the target vascular mask includes the left anterior descending branch mask and / or the diagonal branch mask.

[0178] Specifically, the computer device can determine whether the distance between all bypass candidate points in each target vessel connection domain and the outer surface of the heart chamber is greater than a fifth preset distance threshold. If the judgment result is yes, these bypass candidate points can be regarded as internal mammary artery bypass data. At this time, the total number of bypass candidate points belonging to the left anterior descending branch mask and / or the diagonal branch mask among all vessel segmentation points in the target vessel connection domain can be obtained based on the target vessel mask, that is, the third total number. In addition, the computer device can also determine whether the distance between all bypass candidate points in each target vessel connection domain and the outer surface of the heart chamber is equal to a fifth preset distance threshold. If the judgment result is yes, these bypass candidate points can also be regarded as internal mammary artery bypass data. The above-mentioned fifth preset distance threshold, fourth preset distance threshold, second preset distance threshold and third preset distance threshold can be equal or unequal and can be determined according to actual conditions. In this embodiment, the above-mentioned fifth preset distance threshold can be equal to 5 cm.

[0179] In addition, the computer device can determine whether the closest distance between all candidate bypass points within the target vessel connectivity domain to which the candidate bypass point belongs and the outer surface of the heart chamber is greater than a preset distance. If so, the candidate bypass point can be discarded. The preset distance can be determined based on actual conditions. In this embodiment, the preset distance can be equal to 0.5 cm.

[0180] S2234c: If the third total number is greater than or equal to a fourth preset number threshold, determine the bridging candidate points belonging to the left anterior descending branch mask and the diagonal branch mask among all the vessel segmentation points in the target vessel connectivity domain as second retained candidate points.

[0181] Specifically, the computer device may determine whether the third total number is greater than or equal to a fourth preset number threshold. If so, the computer device may retain the bridging candidate points that belong to the left anterior descending artery mask or the diagonal branch mask among all the vessel segmentation points within the target vessel connectivity domain, and determine these bridging candidate points as second retained candidate points. The fourth preset number threshold may be determined based on actual circumstances.

[0182] S2234d: Determine the first retained candidate point and the second retained candidate point as bridge starting points.

[0183] Furthermore, the computer device may determine the acquired first candidate point and second candidate point as bridge starting points. The number of bridge starting points may be equal to the total number of the first candidate point and the second candidate point.

[0184] The above-mentioned bridge reconstruction method can determine the starting point of the bridge, and then obtain the bridge reconstruction result through processing the starting point of the bridge; this method avoids manual participation, can save human resources and time for determining the starting point of the bridge, and reduce the cost of determining the starting point of the bridge. At the same time, no manual participation is required to determine the starting point of the bridge, which can improve the accuracy of the determined starting point of the bridge, and further improve the accuracy of the bridge reconstruction result; in addition, this method can determine the starting point of the bridge within a smaller range, thereby reducing the amount of calculation for determining the starting point of the bridge, improving the speed of determining the starting point of the bridge, and can quickly obtain the starting point of the bridge in a short time, thereby improving the efficiency of bridge reconstruction.

[0185] As one example, Figure 13 As shown, the step of obtaining the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result in S2200 can be implemented by the following steps:

[0186] S2240: Determine corresponding blood vessel segmentation points according to the blood vessel segmentation result.

[0187] Specifically, the above-mentioned blood vessel segmentation result may be blood vessel segmentation data. The blood vessel segmentation data may be in the form of point coordinates, which are displayed in a coordinate system, and each blood vessel segmentation result has a corresponding blood vessel segmentation point.

[0188] S2250: trace back each blood vessel segmentation point along the bypass path from the end point of the bypass to the starting point of the bypass, and determine the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point.

[0189] Specifically, the computer device may sequentially trace back each vessel segmentation point along each bypass path, starting from the bypass end point of each bypass path toward the bypass start point. Based on the target vessel mask of the traced vessel segmentation point, any traced vessel segmentation point may be determined as an anastomosis between the bypass and normal vessels. The number of determined anastomosis between the bypass and normal vessels may be greater than or equal to one. In this embodiment, the bypass path may be predetermined, and the bypass end point may be the bypass end point in the predetermined bypass path.

[0190] Among them, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S2250 may include: if the bypass tail point is determined to be the blood vessel mask point based on the target blood vessel mask of the blood vessel segmentation point, and the current blood vessel segmentation point is the bypass mask point, backtracking is stopped, and the current blood vessel segmentation point is determined as the anastomosis between the bypass and the normal blood vessel.

[0191] It should be noted that if the backtracked bypass tail point is determined to be a vascular mask point based on the target vascular mask of each vascular segmentation point, and the computer device stops backtracking when the current vascular segmentation point in the bypass path trajectory to which the bypass tail point belongs is a bypass mask point, then the computer device can determine the backtracked current vascular segmentation point as the anastomosis between the bypass and the normal vessel. In this case, the anastomosis between the bypass and the normal vessel can be the bypass starting point or any vascular segmentation point between the bypass tail point and the bypass starting point. The above-mentioned vascular mask point can be understood as the vascular segmentation point where the target vascular mask is the vascular mask; the above-mentioned bypass mask point can be understood as the vascular segmentation point where the target vascular mask is the bypass mask.

[0192] Among them, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S2250 may also include: if the end point of the bypass is a bypass mask point, then backtracking is stopped when the current blood vessel segmentation point is a blood vessel mask point, and the current blood vessel segmentation point is determined as the anastomosis between the bypass and the normal blood vessel.

[0193] It can be understood that if the backtracked bypass tail point is determined to be the bypass mask point based on the target vessel mask of each vessel segmentation point, then the computer device can continue to backtrack forward and stop backtracking when the current vessel segmentation point in the bypass path trajectory to which the bypass tail point belongs is the vessel mask point. At this time, the current vessel segmentation point can be determined as the anastomosis between the bypass and the normal vessel.

[0194] At the same time, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S2250 may also include: if all blood vessel segmentation points between the current blood vessel segmentation point and the bypass tail point are bypass mask points, and the backtracking distance between the current blood vessel segmentation point and the bypass tail point is greater than the preset distance threshold, then the bypass tail point is determined as the anastomosis between the bypass and the normal blood vessel.

[0195] It is understandable that if all the vascular segmentation points (including the current vascular segmentation point and the bypass tail point) between the current vascular segmentation point and the bypass tail point traced back by the computer device are all bypass mask points, and the traced distance between the current vascular segmentation point and the bypass tail point is greater than or equal to the sixth preset distance threshold, then the bypass tail point can be determined as the anastomosis between the bypass and the normal blood vessel. Among them, the current vascular segmentation point traced back and the corresponding bypass tail point are on the same bypass path trajectory. The above-mentioned sixth preset distance threshold can be determined according to actual conditions, and the specific value is not limited. In this embodiment, the above-mentioned sixth preset distance threshold can be equal to 3cm.

[0196] The above-mentioned bypass reconstruction method can determine the anastomosis between the bypass and the normal blood vessel, and then process the anastomosis between the bypass and the normal blood vessel, the starting point of the bypass and the end point of the bypass to obtain the bypass reconstruction result. This method avoids manual reconstruction of the bypass model, can save human resources and bypass reconstruction time, reduce the cost of bypass reconstruction, and further improve the efficiency of bypass reconstruction.

[0197] The third bridge reconstruction method

[0198] The bridge reconstruction method can be described in detail through the following examples:

[0199] like Figure 14 The figure shows a flow chart of a bridge reconstruction method provided by an embodiment, wherein the method is applied to Figure 1 The computer device in the example is used to illustrate the process, including the following steps:

[0200] S3000: Obtain heart segmentation results and blood vessel segmentation results.

[0201] Specifically, the computer device can input multiple frames of cardiac images of the diagnosis and treatment object into a pre-trained cardiac segmentation model to obtain a cardiac segmentation result. Further, the vascular range is determined by the cardiac segmentation result, and the vascular range is segmented to obtain a vascular segmentation result. Optionally, the vascular segmentation result can be a segmentation result corresponding to the entire vascular area in the cardiac segmentation result. The above-mentioned cardiac image can be a three-dimensional CT enhanced image in dicom format. The above-mentioned cardiac segmentation result can be a cardiac segmentation image containing the various chambers of the heart, the aorta, and the aortic arch, and the various chambers of the heart, the aorta, and the aortic arch in the cardiac segmentation image have corresponding labels to distinguish different tissues in the cardiac segmentation image. The various chambers of the heart can be the left atrium, right atrium, left ventricle, and right ventricle of the heart. In addition, the above-mentioned cardiac segmentation result can also be represented by cardiac segmentation data containing the various chambers of the heart, the aorta, and the aortic arch.

[0202] It is understandable that the above-mentioned pre-trained cardiac segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model; wherein, the computer device can perform network model training on the initial cardiac segmentation model through a cardiac image training set to obtain a pre-trained cardiac segmentation model. Specifically, the computer device can input the cardiac image in the cardiac image training set into the initial cardiac segmentation model to obtain a cardiac segmentation prediction result, calculate the prediction error value between the cardiac segmentation prediction result and the standard cardiac segmentation result through a loss function, and update the initial network parameters in the initial cardiac segmentation model according to the prediction error value, and continuously iterate the above training steps until the prediction error value meets the preset error threshold or the number of iterations reaches the preset iteration threshold, thereby obtaining a pre-trained cardiac segmentation model. The above-mentioned cardiac image training set can be a collection of cardiac images of different diagnostic and treatment subjects, and the above-mentioned standard cardiac segmentation result can be an idealized cardiac segmentation result.

[0203] At the same time, the heart segmentation result can be taken as a whole to determine the blood vessel range. The computer device can extend the heart segmentation result outward by a certain preset range to obtain the blood vessel range, and further input the blood vessel range into a pre-trained blood vessel segmentation model to obtain the blood vessel segmentation result. The above preset range can be customized according to actual conditions, as long as the blood vessel range is smaller than the range of the heart image. If the heart segmentation result is represented by a heart segmentation image, the blood vessel range can be an image representation obtained by extending the heart segmentation image outward by a certain preset range; if the heart segmentation result is represented by heart segmentation data, the blood vessel range can be a heart segmentation data representation corresponding to the image obtained by extending the heart segmentation image outward by a certain preset range. The above blood vessel segmentation result can be a binary blood vessel segmentation result.

[0204] The pre-trained vascular segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained vascular segmentation model and the pre-trained cardiac segmentation model have the same structure, their network parameters can be different after network model training. The computer device can perform network model training on the initial vascular segmentation model using a vascular range training set to obtain the pre-trained vascular segmentation model. Specifically, the computer device can input vascular data corresponding to vascular ranges in the vascular range training set into the initial vascular segmentation model to obtain a vascular segmentation prediction result. The computer device can calculate the prediction error between the vascular segmentation prediction result and the standard vascular segmentation result using a loss function, and update the initial network parameters of the initial vascular segmentation model based on the prediction error. The above training steps are continuously iterated until the prediction error meets a preset error threshold or the number of iterations reaches a preset iteration threshold, thereby obtaining the pre-trained vascular segmentation model. The vascular range training set can be a collection of vascular range data corresponding to different diagnostic and treatment subjects, and the standard vascular segmentation result can be an idealized vascular segmentation result.

[0205] In addition, both the heart segmentation model and the blood vessel segmentation model can complete network model training before executing S3000 in this embodiment.

[0206] S3100 , obtaining a target blood vessel mask based on the heart segmentation result, the blood vessel segmentation result, and the segmentation model.

[0207] Specifically, the computer device may process the heart segmentation result and the blood vessel segmentation result to obtain a processed result, and input the processed result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. Alternatively, the computer device may directly input the heart segmentation result and the blood vessel segmentation result into the segmentation model to obtain a multi-vessel mask corresponding to each point in the blood vessel segmentation result, i.e., a target blood vessel mask. In this embodiment, the heart segmentation result and the blood vessel segmentation result may both be understood as images or data.

[0208] It is understood that the segmentation model can be a pre-trained neural network model. The segmentation model can be composed of at least one of a convolutional neural network model, a recurrent neural network model, and an adversarial neural network model. If the pre-trained segmentation model has the same structure as the pre-trained vascular segmentation model and the pre-trained heart segmentation model, the network parameters of the three models can be different after network model training. Furthermore, the segmentation model can complete network model training before executing S3100 in this embodiment.

[0209] S3200 : Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result. The bypass reconstruction information includes a bypass trajectory and an anastomosis between the bypass and the normal blood vessel.

[0210] Specifically, the computer device can perform transformation, comparison, calculation, screening, and / or analysis on the target vessel mask and the vessel segmentation result to obtain bypass reconstruction information, or filter target bypass data from the target vessel mask according to preset conditions, and analyze the target bypass data and the vessel segmentation result to determine the bypass reconstruction information. The target vessel mask can be understood as a multi-vessel label mask. In this embodiment, the target vessel mask can include at least one vessel mask and a bypass mask. The vessel mask can be any of 15 masks, namely, the right coronary artery (RCA) mask, the right posterior descending coronary artery (R-PDA) mask, the left ventricular right posterior branch (R-PLB) mask, the left main coronary artery (LM) mask, the left anterior descending artery (LAD) mask, the diagonal branch (D) mask, the circumflex artery (LCX) mask, the obtuse marginal branch (OM) mask, the left posterior descending artery (L-PDA) mask, the left posterior left ventricular branch (L-PLB) mask, the intermediate branch (RAMUS) mask, the atrial branch (LACX) mask, the acute marginal branch (AM) mask, the anterior septal branch (S) mask, or the sinoatrial node branch (LSN) mask. The above-mentioned bypass reconstruction information may include at least one of the bypass path, the range of the bypass path, the bypass starting point and bypass end point in the bypass path, the bypass path size, and each path point in the bypass path. However, in this embodiment, the bypass reconstruction information may include the bypass trajectory and the anastomosis between the bypass and the normal blood vessel.

[0211] S3300: Process the bridging reconstruction information to obtain a bridging reconstruction result.

[0212] Specifically, the computer device can process the bypass starting point, the obtained bypass path, and the bypass reconstruction information of the anastomosis between the bypass and the normal blood vessels to obtain a three-dimensional heart bypass model diagram, that is, the bypass reconstruction result. In addition, the computer device can also first segment the normal blood vessels to obtain the centerline information of all blood vessels, and then process the centerline information of all blood vessels and the bypass reconstruction information to obtain the bypass reconstruction result. The above-mentioned bypass starting point can be known information, or it can be the starting point in the bypass path. Furthermore, the computer device can output the bypass reconstruction result and display it to the doctor for viewing, so that the doctor can make regular return visits to the diagnosis and treatment subject based on the bypass reconstruction result.

[0213] In the above-mentioned bypass reconstruction method, the computer equipment can obtain the heart segmentation result and the blood vessel segmentation result, obtain the target blood vessel mask according to the heart segmentation result, the blood vessel segmentation result and the segmentation model, obtain the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result, and obtain the bypass reconstruction result through processing through the bypass reconstruction information; this method is based on the neural network model to obtain the target blood vessel mask, and further obtains the bypass reconstruction information based on the target blood vessel mask, and then determines the bypass reconstruction result through reconstruction of the bypass reconstruction information, which can avoid manual reconstruction of the bypass model, save human resources and bypass reconstruction time, reduce bypass reconstruction costs, and further improve bypass reconstruction efficiency; at the same time, this method realizes bypass reconstruction based on the neural network model, which can avoid reconstruction errors caused by manual participation in the reconstruction process, thereby improving the accuracy of bypass reconstruction.

[0214] In some scenarios, the integrity and accuracy of the trajectory map corresponding to the constructed vascular segmentation points will directly affect the integrity and accuracy of the determined bypass path trajectory. Therefore, in order to improve the integrity and accuracy of the determined bypass path trajectory, in this embodiment, Figure 15 As shown, the step of obtaining the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result in S3200 can be implemented by the following steps:

[0215] S3210: Determine the target blood vessel connected domain based on the blood vessel segmentation result and the target blood vessel mask.

[0216] Specifically, the computer device can select any point among all the blood vessel segmentation points corresponding to the blood vessel segmentation result of the entire blood vessel area as a reference point, and obtain a preset number of blood vessel segmentation points around the reference point, and determine the reference point and all the blood vessel segmentation points that are blood vessel masks among the preset number of blood vessel segmentation points around the reference point according to the target blood vessel mask, and use the reference point and all the blood vessel segmentation points that are blood vessel masks among these preset number of blood vessel segmentation points as a target blood vessel connected domain. At the same time, the computer device can traverse each blood vessel segmentation point corresponding to the blood vessel segmentation result, and then determine the target blood vessel connected domain corresponding to each blood vessel segmentation point.

[0217] It should be noted that if the vessel segmentation result corresponds to a three-dimensional image, a preset number of vessel segmentation points plus one vessel segmentation point (i.e., the reference point and a preset number of vessel segmentation points surrounding the reference point) can be selected from all vessel segmentation points corresponding to the vessel segmentation result to determine the target vessel connected domain. Within the target vessel connected domain, any vessel segmentation point must have at least one other vessel segmentation point connected to the selected current vessel segmentation point. In other words, within the same neighborhood (the preset number of surrounding vessel segmentation points, i.e., the preset number of neighborhoods), at least one other vessel segmentation point can be connected to the current vessel segmentation point. The preset number can be any value, as long as any vessel segmentation point can be connected to at least one other vessel segmentation point within its target vessel connected domain.

[0218] S3220 , constructing a trajectory map corresponding to the blood vessel segmentation points using all corresponding blood vessel segmentation results within the target blood vessel connected domain.

[0219] Specifically, the computer device can perform skeletonization on the vessel segmentation results to obtain at least one vessel centerline. This skeletonization can be understood as reducing a binary object to a 1-pixel wide representation. The starting point of the bridge corresponds to the starting point on the vessel centerline. The number of vessel centerlines can be equal to the number of trajectory maps. Each vessel centerline contains multiple vessel segmentation points, each of which has its own target vessel connected domain.

[0220] Specifically, each vessel segmentation result can be understood as a point coordinate, namely, a vessel segmentation point coordinate. The computer device can first establish a dot graph using all vessel segmentation results corresponding to each target vessel connected domain. Then, based on the positional order of the bypass starting points within each target vessel connected domain and all vessel segmentation points on each vessel centerline, the computer device sequentially connects each adjacent vessel segmentation point in the dot graph, starting from the bypass starting points in the dot graph, to obtain at least one trajectory graph. That is, each trajectory graph includes multiple vessel segmentation points and edges between two adjacent vessel segmentation points. Each edge in the trajectory graph can also be part of a bypass path. The number of trajectory graphs corresponding to each target vessel connected domain can be greater than or equal to 1, or can be equal to the number of bypass starting points contained in the corresponding target vessel connected domain. Each edge in the trajectory graph can be understood as a path within the trajectory graph.

[0221] S3230: Starting from the bypass starting point, traverse the adjacent blood vessel segmentation points backward along the trajectory diagram to determine the bypass trajectory.

[0222] It should be noted that the trajectory diagram can be in the form of a straight line or a broken line, and this is not limited to this. In this embodiment, the trajectory diagram can be in the form of a binary tree. Therefore, the vessel segmentation points in the trajectory diagram may or may not be vessel segmentation points in the bypass path. However, in this embodiment, it is possible to traverse some or all of the vessel segmentation points in the trajectory diagram.

[0223] The above-mentioned bypass reconstruction method can determine the target vascular connectivity domain based on the vascular segmentation results and the target vascular mask, and construct a trajectory map corresponding to the vascular segmentation point through all the corresponding vascular segmentation results in the target vascular connectivity domain. Starting from the bypass starting point, the adjacent vascular segmentation points are traversed backward along the trajectory map to determine the bypass path. This process avoids manual construction of the trajectory map corresponding to the vascular segmentation point, and can achieve mapping through machine learning algorithms, thereby improving the accuracy and completeness of the mapping results. It can further traverse a complete and accurate trajectory map, and can also improve the accuracy and completeness of the determined bypass path trajectory, and can provide accurate and complete bypass reconstruction information for determining the bypass reconstruction results. At the same time, this method avoids manual participation in determining the bypass path, which can save human resources and reduce the workload of medical staff.

[0224] As one example, Figure 16 As shown, the step of starting from the bypass starting point and traversing the adjacent blood vessel segmentation points backward along the trajectory graph to determine the bypass path in S3230 can be achieved by the following steps:

[0225] S3231. Starting from the starting point of the bypass, traverse the adjacent blood vessel segmentation points backward along the trajectory diagram to determine the end point of the bypass.

[0226] Specifically, the computer device may start from the bridge starting point in each trajectory image and sequentially traverse the adjacent blood vessel segmentation points backward to determine the bridge end point among all blood vessel segmentation points in the corresponding trajectory image. A trajectory image may include a bridge starting point and one or more bridge end points.

[0227] S3232: Determine a bypass path based on the bypass starting point, the bypass ending point, and the traversed blood vessel segmentation points.

[0228] It should be noted that the computer device can determine the bypass path based on the bypass starting point, the bypass ending point, and all vascular segmentation points traversed between the bypass starting point and the bypass ending point. A trajectory map can include one or more bypass paths. The bypass starting point can be a predetermined bypass point in the bypass path.

[0229] The above-mentioned bypass reconstruction method can start from the bypass starting point, traverse the adjacent blood vessel segmentation points backward along the trajectory diagram, determine the bypass end point among all blood vessel segmentation points, and determine the bypass trajectory based on the bypass starting point, the bypass end point and the traversed blood vessel segmentation points. This method avoids manual participation in traversal to determine the bypass trajectory, can save human resources, and reduce the workload of medical staff.

[0230] As one example, Figure 17 As shown, the step of traversing the adjacent blood vessel segmentation points in sequence backward along the trajectory diagram to determine the end point of the bypass in S3231 can be achieved by the following steps:

[0231] S3231a: If the current traversed blood vessel segmentation point is a first bifurcation point, determine at least one second bifurcation point adjacent to the first bifurcation point.

[0232] Specifically, if the current blood vessel segmentation point traversed by the computer device is a bifurcation point, i.e., a first bifurcation point, it can be determined that there are at least two adjacent blood vessel segmentation points after the current blood vessel segmentation point. Furthermore, all adjacent blood vessel segmentation points after the first bifurcation point can be referred to as second intersection points, or one or a portion of adjacent blood vessel segmentation points after the first bifurcation point can be referred to as second intersection points. The number of first intersection points can be equal to one, and the number of second intersection points can be greater than one.

[0233] It should be noted that there are at least two vascular segmentation points in the neighborhood after the first bifurcation point. Then, for all adjacent vascular segmentation points after the first bifurcation point, if there are only two vascular segmentation points in the neighborhood and they are vascular segmentation points on the edge of the trajectory graph, the vascular segmentation points that have not been traversed can be selected to continue traversing backward. If there are no vascular segmentation points in the neighborhood, the current second bifurcation point can be determined as the end point of the bypass. If the total number of vascular segmentation points in the neighborhood is greater than 2, the current vascular segmentation point that has been traversed is determined as the second bifurcation point.

[0234] S3231b: If there is no bridging mask point among all the blood vessel segmentation points between the first bifurcation point and each second bifurcation point, determine whether there is a main branch vessel mask point among all the blood vessel segmentation points between the first bifurcation point and each second bifurcation point.

[0235] It is understandable that the computer device can determine whether there are any vessel segmentation points belonging to the bypass mask among all the vessel segmentation points between the first bifurcation point and any second bifurcation point. If it is determined that there are no bypass mask points among all the vessel segmentation points between the first bifurcation point and the current second bifurcation point, the computer device can continue to determine whether there are any vessel segmentation points belonging to the main branch vessel mask among all the vessel segmentation points between the first bifurcation point and the current second bifurcation point. The current second intersection point can be any second intersection point among all the second intersection points. The number of bypass mask points among all the vessel segmentation points between the first bifurcation point and the current second bifurcation point can be greater than or equal to 0, and the number of main branch vessel mask points among all the vessel segmentation points between the first bifurcation point and the current second bifurcation point can be greater than or equal to 0.

[0236] It should be noted that the bypass mask point can be understood as a vessel segmentation point belonging to the bypass mask. The main branch mask point can be understood as a vessel segmentation point belonging to the main branch mask. One or more other vessel segmentation points may exist between the first intersection point and the second intersection point, or there may not be any other vessel segmentation points between the first intersection point and the second intersection point.

[0237] S3231c: If yes, then starting from each second bifurcation point, continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph until the last blood vessel segmentation point in the trajectory graph is traversed, and determine the last blood vessel segmentation point as the end point of the bypass.

[0238] Specifically, if the computer device determines that a main branch vessel mask point exists among all the vessel segmentation points between the first bifurcation point and the current second bifurcation point, the computer device may start from each current second bifurcation point and continue to traverse the adjacent vessel segmentation points backward along the trajectory graph until the last vessel segmentation point on a different path in the trajectory graph is reached, and the last vessel segmentation point is determined as the bridge tail point. The number of bridge tail points may be greater than one.

[0239] It should be noted that if there are intersections after the current second bifurcation point in the trajectory graph, in this case the current second intersection can be used as the first bifurcation point, and the traversed intersections can be used as the current second bifurcation point. The steps in S3231b and S3231c above are continued to be executed until the last blood vessel segmentation point in the trajectory graph is traversed to obtain the end point of the bridge.

[0240] Furthermore, the computer device can determine a bypass path based on each bypass starting point, the corresponding bypass ending point, and all vascular segmentation points traversed between each bypass starting point and the corresponding bypass ending point. The number of bypass paths can be greater than one, and different bypass paths can share a bypass starting point but different bypass ending points.

[0241] Among them, all the traversed blood vessel segmentation points between the bypass starting point and the bypass end point may be blood vessel segmentation points in the bypass path, or may not be blood vessel segmentation points in the bypass path. Figure 18 As shown, the starting point of the bridge in a trajectory diagram is point A, the intersection points after point A are point B and point C, the adjacent blood vessel segmentation points after point B are point D and point E, and the adjacent blood vessel segmentation points after point C are point F and point G. If there is no main branch vessel mask point among all the blood vessel segmentation points traversed between point A and point B, and there is a main branch vessel mask point among all the blood vessel segmentation points traversed between point A and point C, then point C can be retained, point B can be filtered out, and the traversal can be continued from point C backward. The final bridge path trajectory can be path ACF and / or path ACG, that is, the bridge path trajectory does not contain the filtered blood vessel segmentation points and the untraversed blood vessel segmentation points.

[0242] The above-mentioned bypass reconstruction method can determine the bypass tail point, and then obtain the bypass path trajectory through the bypass tail point, and obtain the bypass reconstruction result through the bypass path trajectory and the anastomosis between the bypass and the normal blood vessel. This method avoids manual reconstruction of the bypass model, can save human resources and bypass reconstruction time, reduce bypass reconstruction costs, and further improve bypass reconstruction efficiency.

[0243] In some scenarios, if there are multiple second bifurcation points in the trajectory graph, and there are intersections after each second intersection point, then after the above step S3231a is executed, if Figure 19 As shown, the above-mentioned bridge reconstruction method may further include the following steps:

[0244] S3231d: If there are bridging mask points in all the blood vessel segmentation points between the first intersection point and each second intersection point, then starting from each second bifurcation point, continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph, and determine at least one next traversed intersection point as the third intersection point.

[0245] Specifically, the computer device may determine whether any of the vessel segmentation points between the first bifurcation point and the current second bifurcation point contains a vessel segmentation point that is a bridging mask. If it is determined that any of the vessel segmentation points between the first bifurcation point and the current second bifurcation point contains a bridging mask point, the computer device may traverse the adjacent vessel segmentation points backward along the trajectory graph starting from the current second bifurcation point, and determine at least one next traversed intersection point as a third intersection point. The number of the third intersection points may be greater than or equal to one.

[0246] In addition, if starting from the current second bifurcation point, the adjacent vascular segmentation points are continued to be traversed backward along the trajectory diagram, and it is determined that there is no intersection among the traversed adjacent vascular segmentation points, the total number of bridging mask points in the traversed adjacent vascular segmentation points can be obtained, and it can be determined whether the total number of bridging mask points is less than or equal to the preset number. If the total number of bridging mask points is less than or equal to the preset number, the current second bifurcation point can be discarded, and then starting from other second bifurcation points, the adjacent vascular segmentation points can be traversed backward to determine the third intersection point.

[0247] S3231e: If there are bridging mask points among all the blood vessel segmentation points between the second intersection point and each third intersection point, determine the direction vector angle between the first intersection point and each second intersection point.

[0248] Specifically, the computer device may determine whether any of all the vessel segmentation points between the second intersection point and any third intersection point contains a vessel segmentation point that is a bridging mask. If it is determined that any of the vessel segmentation points between the second intersection point and the current third intersection point contains a bridging mask point, the computer device may calculate a direction vector angle between the first intersection point and the second intersection point immediately preceding the current third intersection point. The current third intersection point may be understood as any third intersection point among all the third intersection points.

[0249] It should be noted that there may be one or more non-intersection points between the second intersection point and the current third intersection point, or there may be no intersection point. The first intersection point and each second intersection point have a corresponding direction vector angle. Starting from the first intersection point, all blood vessel segmentation points between the first intersection point and the current second intersection point (including the first intersection point and the current second intersection point) are numbered in sequence as blood vessel segmentation point 1, blood vessel segmentation point 2, blood vessel segmentation point 3, ..., blood vessel segmentation point n, where the first intersection point is the first blood vessel segmentation point 1 and the last blood vessel segmentation point n is the current second intersection point. The direction vector angle between the first intersection point and the current second intersection point can be equal to the sum of the coordinate value of the first blood vessel segmentation point 1 between the first intersection point and the current second intersection point minus the coordinate value of the second blood vessel segmentation point 2, the coordinate value of the second blood vessel segmentation point 2 minus the coordinate value of the first blood vessel segmentation point 1, the coordinate value of the third blood vessel segmentation point 3 minus the coordinate value of the first blood vessel segmentation point 1, ..., and the coordinate value of the last blood vessel segmentation point n minus the coordinate value of the first blood vessel segmentation point 1, and then normalized to obtain a normalized result, and then the cosine value of the normalized result is calculated, i.e., the direction vector angle. The above-mentioned blood vessel segmentation points can be three-dimensional coordinate points, and the normalized result can be a direction vector.

[0250] S3231f. Determine the target second bifurcation point in the bridge path according to the direction vector angle.

[0251] It should be noted that the computer device may compare the direction vector angles corresponding to the second intersection points, determine the minimum direction vector angle, and determine the second intersection point corresponding to the minimum direction vector angle as the target second bifurcation point in the bridge path. The number of target second bifurcation points determined may be greater than or equal to one.

[0252] At the same time, after the above-mentioned step S3231d is executed, the above-mentioned bypass reconstruction method may also include: if there is a branch mask point in all the blood vessel segmentation points between the first bifurcation point and each second bifurcation point, then the target second bifurcation point in the bypass path is determined based on the maximum number of belonging mask points corresponding to all the blood vessel segmentation points between the first bifurcation point and the previous blood vessel segmentation point that has been traversed.

[0253] It is understood that the computer device can determine whether there is a blood vessel segmentation point belonging to the branch mask among all the blood vessel segmentation points between the first bifurcation point and the current second bifurcation point. If it is determined that there is a branch mask point among all the blood vessel segmentation points between the first bifurcation point and the current second bifurcation point, it can determine whether there is the most attribution mask points between the first bifurcation point and the current second bifurcation point based on the most attribution mask points corresponding to all the blood vessel segmentation points between the first bifurcation point and the previous traversed blood vessel segmentation point. If so, the current second intersection point is used as the target second bifurcation point in the bypass path. The previous traversed blood vessel segmentation point can be the blood vessel segmentation point before the first intersection point.

[0254] Among them, if all the vascular segmentation points between the first bifurcation point and the last traversed vascular segmentation point include 7 vascular segmentation points, namely vascular segmentation point 1 (belonging to the left anterior descending branch mask point), vascular segmentation point 2 (belonging to the left anterior descending branch mask point), vascular segmentation point 3 (belonging to the diagonal branch mask point), vascular segmentation point 4 (belonging to the left anterior descending branch mask point), vascular segmentation point 5 (belonging to the anterior septal branch mask point), vascular segmentation point 6 (belonging to the diagonal branch mask point), and vascular segmentation point 7 (belonging to the anterior septal branch mask point), among which there are 3 left anterior descending branch mask points, 2 anterior septal branch mask points, and 2 diagonal branch mask points. Since the left anterior descending branch belongs to the three main branches, the most belonging mask points can be the main branch mask points.

[0255] S3231g: Starting from the second target intersection point, continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph until the last blood vessel segmentation point in the trajectory graph is traversed, and determine the last blood vessel segmentation point as the end point of the bypass.

[0256] Specifically, the computer device may start from any target second bifurcation point and continue to traverse the blood vessel segmentation points adjacent to the current target second bifurcation point backward along the trajectory graph until it reaches the last blood vessel segmentation point on a different path in the trajectory graph, and determine the last blood vessel segmentation point as the end point of the bypass. The current target second bifurcation point may be any target second intersection point among all target second intersection points.

[0257] The above-mentioned bridge reconstruction method can determine the bridge tail point, and then reconstruct the target bridge data according to the bridge tail point to obtain the bridge reconstruction result. This method avoids manual reconstruction of the bridge model, can save human resources and bridge reconstruction time, reduce bridge reconstruction costs, and further improve bridge reconstruction efficiency.

[0258] As one example, Figure 20 As shown, the step of obtaining the bypass reconstruction information based on the target vessel mask and the vessel segmentation result in S3200 can be implemented by the following steps:

[0259] S3240: Determine corresponding blood vessel segmentation points according to the blood vessel segmentation result.

[0260] Specifically, the above-mentioned blood vessel segmentation result may be blood vessel segmentation data. The blood vessel segmentation data may be in the form of point coordinates, which are displayed in a coordinate system, and each blood vessel segmentation result has a corresponding blood vessel segmentation point.

[0261] S3250: trace back each blood vessel segmentation point along the bypass path from the end point of the bypass to the starting point of the bypass, and determine the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point.

[0262] Specifically, the computer device may trace back each vessel segmentation point along each bypass path, starting from the end of each bypass path toward the start of each bypass path. Based on the target vessel mask of the traced vessel segmentation point, any traced vessel segmentation point may be determined as an anastomosis between the bypass and the normal vessel. The number of determined anastomosis between the bypass and the normal vessel may be greater than or equal to one.

[0263] Among them, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S3250 may include: if the end point of the bypass is a blood vessel mask point, and the current blood vessel segmentation point traced back is the bypass mask point, stopping the backtracking and determining the current blood vessel segmentation point as the anastomosis between the bypass and the normal blood vessel.

[0264] It should be noted that if the backtracked bypass end point is determined to be a vessel mask point based on the target vessel mask of each vessel segmentation point, and the computer device stops backtracking when the current vessel segmentation point in the bypass path to which the bypass end point belongs is a bypass mask point, the computer device can then determine the current vessel segmentation point to be the anastomosis between the bypass and the normal vessel. The anastomosis between the bypass and the normal vessel can be a bifurcation point, that is, any intersection between the bypass end point and the bypass starting point.

[0265] In addition, in this embodiment, the vascular mask of the edge between two adjacent vascular segmentation points can be determined based on the target vascular mask corresponding to each vascular segmentation point. The vascular mask of the edge between two adjacent vascular segmentation points can be the target vascular mask corresponding to the most attributed mask points among all vascular segmentation points between the two adjacent vascular segmentation points. The vascular mask of each edge can be a vascular mask or a bridging mask.

[0266] If the current vessel segmentation point (i.e., bifurcation point) is the end point of the bypass bridge of the vessel mask, and the previous edge connected to the current vessel segmentation point is the bypass bridge mask, the current vessel segmentation point can be determined to be the anastomosis between the bypass bridge and the normal vessel. If the current vessel segmentation point is the end point of the bypass bridge mask (i.e., bifurcation point), and the current vessel segmentation point contains the vessel mask, the current vessel segmentation point is determined to be the anastomosis between the bypass bridge and the normal vessel.

[0267] Among them, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S3250 may also include: if the end point of the bypass is a bypass mask point, then backtracking is stopped when the current blood vessel segmentation point is a blood vessel mask point, and the current blood vessel segmentation point is determined as the anastomosis between the bypass and the normal blood vessel.

[0268] It can be understood that if the backtracked bypass tail point is determined to be the bypass mask point based on the target vessel mask of each vessel segmentation point, then the computer device can continue to backtrack forward and stop backtracking when the current vessel segmentation point in the bypass path trajectory to which the bypass tail point belongs is the vessel mask point. At this time, the current vessel segmentation point can be determined as the anastomosis between the bypass and the normal vessel.

[0269] At the same time, the step of determining the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point in the above S3250 may also include: if all blood vessel segmentation points between the current blood vessel segmentation point and the bypass tail point are bypass mask points, and the backtracking distance between the current blood vessel segmentation point and the bypass tail point is greater than the preset distance threshold, then the bypass tail point is determined as the anastomosis between the bypass and the normal blood vessel.

[0270] It is understandable that if all the vascular segmentation points (including the current vascular segmentation point and the bypass tail point) between the current vascular segmentation point and the bypass tail point traced back by the computer device are all bypass mask points, and the traced distance between the current vascular segmentation point and the bypass tail point is greater than or equal to the preset distance threshold, then the bypass tail point can be determined as the anastomosis between the bypass and the normal blood vessel. Among them, the current vascular segmentation point traced back and the corresponding bypass tail point are on the same bypass path trajectory. The above-mentioned preset distance threshold can be determined according to actual conditions, and the specific value is not limited. In this embodiment, the above-mentioned preset distance threshold can be equal to 3cm.

[0271] The above-mentioned bypass reconstruction method can determine the anastomosis between the bypass and the normal blood vessel, and then process the anastomosis between the bypass and the normal blood vessel, the starting point of the bypass and the end point of the bypass to obtain the bypass reconstruction result. This method avoids manual reconstruction of the bypass model, can save human resources and bypass reconstruction time, reduce the cost of bypass reconstruction, and further improve the efficiency of bypass reconstruction.

[0272] It should be understood that although Figure 1-20 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. In addition, Figure 1-20 At least part of the steps may include multiple steps or multiple stages. These steps or stages are not necessarily performed at the same time, but can be performed at different times. The order of execution of these steps or stages is not necessarily one by one, but can be performed in turn or alternately with other steps or at least part of the steps or stages in other steps.

[0273] In one embodiment, Figure 21 As shown, a bridge reconstruction device is provided, comprising: a segmentation result acquisition module 31, a mask acquisition module 32, a bridge information acquisition module 33 and a reconstruction module 34, wherein:

[0274] A segmentation result acquisition module 31 is used to obtain heart segmentation results and blood vessel segmentation results;

[0275] a mask acquisition module 32 for acquiring a target blood vessel mask based on the heart segmentation result, the blood vessel segmentation result, and the segmentation model;

[0276] A bypass information acquisition module 33 is configured to obtain bypass reconstruction information based on the target vessel mask and the vessel segmentation result. The bypass reconstruction information includes a bypass trajectory and an anastomosis between the bypass and the normal vessel.

[0277] The reconstruction module 34 is configured to process the bridging reconstruction information to obtain a bridging reconstruction result.

[0278] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0279] In one embodiment, the bridging information acquisition module 33 includes: a connected domain determination unit, a trajectory map construction unit, and a behavior trajectory determination unit, wherein:

[0280] a connected domain determining unit, configured to determine a target blood vessel connected domain based on the blood vessel segmentation result and the target blood vessel mask;

[0281] A trajectory map construction unit is used to construct a trajectory map corresponding to the vascular segmentation point through all corresponding vascular segmentation results in the target vascular connectivity domain;

[0282] The path trajectory determination unit is used to start from the bypass starting point and traverse the adjacent blood vessel segmentation points backward along the trajectory diagram to determine the bypass path.

[0283] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0284] In one embodiment, the movement trajectory determination unit includes: an end point determination subunit and a movement trajectory determination subunit, wherein:

[0285] The tail point determination subunit is used to start from the starting point of the bypass and traverse the adjacent blood vessel segmentation points backward along the trajectory graph to determine the tail point of the bypass;

[0286] The path trajectory determination subunit is used to determine the path trajectory of the bypass according to the starting point of the bypass, the end point of the bypass and all the traversed blood vessel segmentation points.

[0287] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0288] In one embodiment, the tail point determination subunit includes: a first determination subunit, a judgment subunit, and a second determination subunit, wherein:

[0289] a first determining subunit, configured to, when the current traversed blood vessel segmentation point is a first bifurcation point, determine at least one second bifurcation point adjacent to the first bifurcation point;

[0290] a judgment subunit, configured to judge whether there is a main branch vessel mask point among all the vessel segmentation points between the first bifurcation point and each second bifurcation point when no bridging mask point exists among all the vessel segmentation points between the first bifurcation point and each second bifurcation point;

[0291] The second determining subunit is configured to, when the judgment result of the judging subunit is yes, start from each second bifurcation point and continue to traverse the adjacent blood vessel segmentation points backward along the trajectory diagram until the last blood vessel segmentation point in the trajectory diagram is traversed, and determine the last blood vessel segmentation point as the end point of the bypass.

[0292] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0293] In one embodiment, the tail point determination subunit further includes: a third determination subunit, a vector angle determination subunit, a fourth determination subunit, and a loop execution subunit, wherein:

[0294] a third determining subunit, configured to, when a bridging mask point exists in all the blood vessel segmentation points between the first intersection point and each second intersection point, start from each second bifurcation point, continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph, and determine at least one next traversed intersection point as a third intersection point;

[0295] a vector angle determination subunit, configured to determine a direction vector angle between the first intersection point and each second intersection point when a bridging mask point exists in all the blood vessel segmentation points between the second intersection point and each third intersection point;

[0296] a fourth determining subunit, configured to determine a target second bifurcation point in the bridge path according to the direction vector angle;

[0297] The loop execution subunit is used to start from the second target intersection point and continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph until the last blood vessel segmentation point in the trajectory graph is traversed, and the last blood vessel segmentation point is determined as the end point of the bridge.

[0298] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0299] In one embodiment, the tail point determination subunit further includes: a fifth determination subunit, wherein:

[0300] The fifth determination subunit is configured to determine the target second bifurcation point in the bypass path according to the maximum number of attribution mask points corresponding to all the vessel segmentation points between the first bifurcation point and the previous vessel segmentation point traversed, when there is a branch mask point among all the vessel segmentation points between the first bifurcation point and each second bifurcation point.

[0301] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0302] In one embodiment, the bridging information acquisition module 33 includes: a segmentation point determination unit and an anastomotic stoma determination unit, wherein:

[0303] a segmentation point determination unit, configured to determine corresponding blood vessel segmentation points according to the blood vessel segmentation result;

[0304] The anastomosis determination unit is used to trace back each blood vessel segmentation point along the bypass path from the end point of the bypass to the starting point of the bypass, and determine the anastomosis between the bypass and the normal blood vessel based on the target blood vessel mask of the blood vessel segmentation point.

[0305] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0306] In one embodiment, the anastomotic stoma determination unit includes: a sixth determination subunit, wherein:

[0307] The sixth determining subunit is configured to stop backtracking when the end point of the bypass is a blood vessel mask point and the current blood vessel segmentation point is a bypass mask point, and determine the current blood vessel segmentation point as the anastomosis between the bypass and the normal blood vessel.

[0308] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0309] In one embodiment, the anastomotic stoma determination unit further includes: a seventh determination subunit, wherein:

[0310] The seventh determining subunit is configured to stop tracing back when the current blood vessel segmentation point is a blood vessel mask point when the bypass tail point is a bypass mask point, and determine the current blood vessel segmentation point as the anastomosis between the bypass and the normal blood vessel.

[0311] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0312] In one embodiment, the anastomotic stoma determination unit further includes: an eighth determination subunit, wherein:

[0313] The eighth determination subunit is used to determine the bypass tail point as the anastomosis between the bypass and the normal blood vessel when all blood vessel segmentation points between the current blood vessel segmentation point and the bypass tail point are bypass mask points and the backtracking distance between the current blood vessel segmentation point and the bypass tail point is greater than a preset distance threshold.

[0314] The bridge reconstruction device provided in this embodiment can execute the above method embodiment. Its implementation principle and technical effects are similar and will not be described in detail here.

[0315] The specific definition of the bypass reconstruction device can be found in the definition of the bypass reconstruction method above and will not be repeated here. Each module in the aforementioned bypass reconstruction device can be implemented in whole or in part via software, hardware, or a combination thereof. Each of the aforementioned modules can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0316] In one embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0317] Obtain heart segmentation results and blood vessel segmentation results;

[0318] Obtain target vessel mask based on heart segmentation results, vessel segmentation results and segmentation model;

[0319] Based on the target vessel mask and the vessel segmentation result, the bypass reconstruction information is obtained, which includes the bypass trajectory and the anastomosis between the bypass and the normal vessel;

[0320] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0321] In one embodiment, a readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0322] Obtain heart segmentation results and blood vessel segmentation results;

[0323] Obtain target vessel mask based on heart segmentation results, vessel segmentation results and segmentation model;

[0324] Based on the target vessel mask and the vessel segmentation result, the bypass reconstruction information is obtained, which includes the bypass trajectory and the anastomosis between the bypass and the normal vessel;

[0325] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0326] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0327] Obtain heart segmentation results and blood vessel segmentation results;

[0328] Obtain target vessel mask based on heart segmentation results, vessel segmentation results and segmentation model;

[0329] Based on the target vessel mask and the vessel segmentation result, the bypass reconstruction information is obtained, which includes the bypass trajectory and the anastomosis between the bypass and the normal vessel;

[0330] The bridge reconstruction information is processed to obtain a bridge reconstruction result.

[0331] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0332] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, 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.

[0333] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A bridge reconstruction method, characterized in that: The method comprises: Obtain heart segmentation results and blood vessel segmentation results; Acquire, based on the heart segmentation result, the blood vessel segmentation result, and the segmentation model, a multi-vessel mask corresponding to each blood vessel segmentation point in the blood vessel segmentation result, i.e., a target blood vessel mask; the target blood vessel mask includes a bypass mask and at least one blood vessel mask; Obtaining bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result, the bypass reconstruction information including a bypass trajectory and an anastomosis between the bypass and the normal blood vessel; Processing the bridge reconstruction information to obtain a bridge reconstruction result; The step of obtaining the bypass reconstruction information based on the target blood vessel mask and the blood vessel segmentation result includes: determining a target blood vessel connected domain according to the blood vessel segmentation result and the target blood vessel mask; Constructing a trajectory graph corresponding to each vessel segmentation point on at least one vessel centerline using all vessel segmentation results corresponding to the target vessel connected domain; wherein all vessel segmentation points in the trajectory graph include a bridge starting point and at least one bridge ending point; and wherein the trajectory graph includes a graph formed by each vessel segmentation point and paths between any two adjacent vessel segmentation points; Starting from the bridging starting point in the trajectory diagram, the adjacent blood vessel segmentation points are traversed backward along the trajectory diagram to determine the bridging path trajectory in the bridging reconstruction information; the backward direction indicates the extension direction from the bridging starting point to the blood vessel segmentation point adjacent to the bridging starting point in the trajectory diagram.

2. The method according to claim 1, characterized in that Starting from the bypass starting point in the trajectory map, traversing the adjacent blood vessel segmentation points in sequence backward along the trajectory map to determine the bypass path in the bypass reconstruction information, including: Starting from the bypass starting point, traversing the adjacent blood vessel segmentation points backward along the trajectory diagram to determine the bypass end point; The bypass path is determined according to the bypass starting point, the bypass ending point, and the traversed blood vessel segmentation points.

3. The method according to claim 2, characterized in that Traversing the adjacent blood vessel segmentation points in sequence backward along the trajectory graph to determine the bridge tail points among all blood vessel segmentation points, including: If the current traversed blood vessel segmentation point is a bifurcation point, that is, a first bifurcation point, then determining at least one second bifurcation point adjacent to the first bifurcation point; If no bypass mask point exists among all the vessel segmentation points between the first bifurcation point and each second bifurcation point, determining whether there is a main branch vessel mask point among all the vessel segmentation points between the first bifurcation point and each second bifurcation point; the bypass mask point indicates a vessel segmentation point where the target vessel mask is a bypass mask; the main branch vessel mask point indicates a vessel segmentation point where the target vessel mask is a main branch vessel mask; If so, starting from each second bifurcation point, the adjacent blood vessel segmentation points are traversed backward along the trajectory graph until the last blood vessel segmentation point in the trajectory graph is traversed, and the last blood vessel segmentation point is determined as the end point of the bypass.

4. The method according to claim 3, characterized in that The method further comprises: If there are bridging mask points among all the blood vessel segmentation points between the first bifurcation point and each of the second bifurcation points, then starting from each of the second bifurcation points, continue to traverse the adjacent blood vessel segmentation points backward along the trajectory graph, and determine at least one next bifurcation point traversed as a third bifurcation point; If there are bridging mask points among all the blood vessel segmentation points between the second bifurcation point and each of the third bifurcation points, determining a direction vector angle between the first bifurcation point and each of the second bifurcation points; Determining the second bifurcation point corresponding to the minimum direction vector angle among the direction vector angles as the target second bifurcation point in the bridge trajectory; Starting from the target second bifurcation point, the adjacent blood vessel segmentation points are traversed in the backward direction along the trajectory graph until the last blood vessel segmentation point in the trajectory graph is traversed, and the last blood vessel segmentation point is determined as the end point of the bypass.

5. The method according to claim 4, characterized in that The method further comprises: If a branch mask point exists among all the vessel segmentation points between the first bifurcation point and each of the second bifurcation points, then based on the most attribution mask points corresponding to all the vessel segmentation points between the first bifurcation point and the previous traversed vessel segmentation point, it is determined whether the most attribution mask points exist between the first bifurcation point and the current second bifurcation point. If so, the current second bifurcation point is determined as the target second bifurcation point in the bypass path; the branch mask point indicates that the target vessel mask is a vessel segmentation point with a branch mask.

6. The method according to claim 1, wherein Obtaining bypass reconstruction information based on the target vessel mask and the vessel segmentation result, including: Determining corresponding blood vessel segmentation points according to the blood vessel segmentation result; Along the bypass path, each vascular segmentation point is traced back in sequence from the bypass end point of the bypass path to the bypass starting point of the bypass path. If the bypass end point is a vascular mask point and the current vascular segmentation point traced back is a bypass mask point, the backtracking is stopped, and the current vascular segmentation point is determined as the anastomosis between the bypass and the normal blood vessel; the vascular mask point indicates the vascular segmentation point where the target vascular mask is the vascular mask, and the bypass mask point indicates the vascular segmentation point where the target vascular mask is the bypass mask.

7. The method according to claim 6, characterized in that The method further comprises: If all the vessel segmentation points between the current vessel segmentation point and the bypass tail point are the bypass mask points, and the backtracking distance between the current vessel segmentation point and the bypass tail point is greater than the preset distance threshold, then the bypass tail point is determined as the anastomosis between the bypass and the normal vessel.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

9. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images