Puncture path generation method and device, equipment and storage medium

Abstract

The invention relates to the technical field of medical treatment, and discloses a puncture path generation method and device, equipment and a storage medium. Firstly, a liver multi-energy-spectrum enhanced image sequence including a multi-energy-spectrum image sequence, a vein phase image sequence and an arterial phase image sequence is obtained. Then, performing image segmentation on the liver multi-energy-spectrum enhanced image sequence to obtain a target key region comprising a first target key region associated with the multi-energy-spectrum image sequence, a second target key region in the vein phase image sequence and a third target key region in the arterial phase image sequence; then, based on the first target key area and the second target key area, determining an expansion penetration area and a puncture target area; in the process, manual intervention is reduced, and the accuracy of determining the target puncture path is improved. And finally, based on the expansion penetration area, the puncture target area and the target key area, a target puncture path is generated, the puncture safety and the puncture precision are improved, and a doctor is assisted to carry out more accurate preoperative planning.

Description

Technical Field

[0001] The present invention relates to the field of medical technology, and in particular, to a method, apparatus, device, and storage medium for generating a puncture path. Background Art

[0002] Percutaneous Transhepatic Cholangial Drainage (PTCD) is a widely used treatment technique in biliary surgery, and is commonly used for diseases such as cholangitis and cholangiocarcinoma.

[0003] In the related art, median filtering, edge detection, and Hough transform in image processing methods are used to obtain the center point of the key area, and virtual reality guidance puncture is performed in combination with three-dimensional data.

[0004] However, the image processing method in the related art has poor generalization ability and inaccurate center point extraction. Therefore, a new puncture path generation method needs to be proposed. Summary of the Invention

[0005] The embodiments of the present specification aim to solve at least one of the technical problems in the related art to some extent. For this reason, the embodiments of the present specification propose a method, apparatus, device, and storage medium for generating a puncture path.

[0006] The embodiments of the present specification provide a method for generating a puncture path, the method including:

[0007] Obtaining a liver multi-energy spectrum enhanced image sequence, wherein the liver multi-energy spectrum enhanced image includes a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence;

[0008] Performing image segmentation on the liver multi-energy spectrum enhanced image sequence to obtain a target key area, wherein the target key area includes a first target key area associated with the multi-energy spectrum image sequence, a second target key area in the venous phase image sequence, and a third target key area in the arterial phase image sequence;

[0009] Determining an expansion penetration area and a puncture target area based on the first target key area and the second target key area;

[0010] Generating a target puncture path based on the expansion penetration area, the puncture target area, and the target key area.

[0011] In one of the embodiments, the generating a target puncture path based on the expansion penetration area, the puncture target area, and the target key area includes:

[0012] Using the puncture path constraint conditions, based on the dilated penetration region, the puncture target region, and the target key region, determine a set of candidate puncture paths;

[0013] Using a distance calculation method, determine the target puncture path from the set of candidate puncture paths.

[0014] In one embodiment, the puncture path constraint conditions include:

[0015] The puncture path does not pass through the hepatic artery region, the hepatic vein region, the portal vein region, or the rib region;

[0016] The puncture angle corresponding to the puncture path is less than a preset constraint angle;

[0017] The puncture path does not repeatedly pass through the same branch bile duct;

[0018] The puncture path does not pass through intrahepatic bile duct stones and bile duct tumors.

[0019] In one embodiment, the multi-energy spectrum image sequence includes a plain scan image sequence and an energy spectrum image sequence, the intrahepatic bile duct stones include positive stones and negative stones, and the method further includes determining the intrahepatic bile duct stone region and the bile duct tumor region by the following means:

[0020] Using a preset density to segment the bile duct region in the plain scan image sequence to determine the positive stone region;

[0021] Based on the bile duct region in the multi-energy spectrum image sequence, determine the energy spectrum curve;

[0022] Based on the slope of the energy spectrum curve, determine the negative stone and the bile duct tumor region.

[0023] In one embodiment, the obtaining of the multi-energy spectrum enhanced liver image sequence includes:

[0024] Obtain an initial multi-energy spectrum enhanced liver image sequence, where the initial multi-energy spectrum enhanced liver image sequence includes the venous phase image sequence, the initial multi-energy spectrum image sequence, and the initial arterial phase image sequence;

[0025] Based on the venous phase image sequence, perform rigid registration on the initial multi-energy spectrum image sequence and the initial arterial phase image sequence to obtain the multi-energy spectrum enhanced liver image sequence.

[0026] In one embodiment, obtaining the initial multi-energy spectrum image sequence by photon CT includes:

[0027] Set multiple energy thresholds;

[0028] Collect photon counting data at each energy threshold and perform image reconstruction to generate the initial multi - energy spectrum image sequence.

[0029] In one embodiment, the multi - energy spectrum image sequence includes a non - contrast image sequence. Image segmentation of the liver multi - energy spectrum enhanced image sequence to obtain the target key regions includes:

[0030] Use a non - contrast segmentation model to perform image segmentation on the non - contrast image sequence to obtain the first target key region;

[0031] Use a venous phase segmentation model to perform image segmentation on the venous phase image sequence to obtain the second target key region;

[0032] Use an arterial phase segmentation model to perform image segmentation on the arterial phase image sequence to obtain the third target key region.

[0033] In one embodiment, the first target key region includes the xiphoid process region, rib region, and liver region. The penetration region includes a left penetration region and a right penetration region. Based on the first target key region and the second target key region, determining the penetration region and the puncture target region includes:

[0034] Perform mapping based on the xiphoid process region to determine the left penetration region;

[0035] Perform mapping based on the rib region and liver region to determine the right penetration region;

[0036] Perform screening and partitioning based on the second target key region to determine the puncture target region.

[0037] In one embodiment, the second target key region includes the bile duct region. Based on the second target key region, performing screening and partitioning to determine the puncture target region includes:

[0038] Extract the centerline corresponding to the bile duct region, where the centerline corresponds to a bifurcation center point;

[0039] Determine a puncture target candidate set from the bile duct region based on a preset screening condition and the centerline;

[0040] Partition the puncture target candidate set into a left puncture target region and a right puncture target region based on the bifurcation center point.

[0041] An embodiment of this specification provides a puncture path generation device, and the device includes:

[0042] A multi-spectrum enhanced image acquisition module for acquiring a sequence of multi-spectrum enhanced liver images, wherein the multi-spectrum enhanced liver images include a multi-spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence;

[0043] A target key region segmentation module for segmenting the sequence of multi-spectrum enhanced liver images to obtain a target key region, wherein the target key region includes a first target key region associated with the multi-spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence;

[0044] An insertion and target point region determination module for determining an expansion insertion region and a puncture target region based on the first target key region and the second target key region;

[0045] A target puncture path generation module for generating a target puncture path based on the expansion insertion region, the puncture target region, and the target key region.

[0046] An embodiment of this specification provides a medical imaging device, which includes: a memory, and one or more processors communicatively connected to the memory; instructions executable by the one or more processors are stored in the memory, and when the instructions are executed by the one or more processors, the one or more processors are caused to implement the steps of the method described in any one of the above embodiments.

[0047] An embodiment of this specification provides a computer device, which includes: a memory, and one or more processors communicatively connected to the memory; instructions executable by the one or more processors are stored in the memory, and when the instructions are executed by the one or more processors, the one or more processors are caused to implement the steps of the method described in any one of the above embodiments.

[0048] An embodiment of this specification provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the above embodiments are implemented.

[0049] An embodiment of this specification provides a computer program product, which includes instructions that, when executed by a processor of a computer device, enable the computer device to execute the steps of the method described in any one of the above embodiments.

[0050] In the above-described embodiment of the specification, first, a liver multi-energy spectrum enhanced image sequence including a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence is obtained. Next, image segmentation is performed on the liver multi-energy spectrum enhanced image sequence to obtain target key regions including a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence. Then, based on the first target key region and the second target key region, a dilated penetration region and a puncture target region are determined. This process can reduce manual intervention, thereby improving the accuracy of determining the target puncture path. Finally, based on the dilated penetration region, the puncture target region, and the target key regions, a target puncture path is generated. This embodiment determines the target puncture path based on the liver multi-energy spectrum enhanced image sequence, improves puncture safety and puncture accuracy, assists the doctor in performing a more accurate preoperative plan, and further provides more efficient and accurate surgical operation support for the doctor, reducing the risks during the operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1a is a schematic flowchart of the puncture path generation method provided by the embodiment of the present specification;

[0052] Figure 1b is a schematic diagram of the target key region provided by the embodiment of the present specification;

[0053] Figure 2a is a schematic flowchart of generating the target puncture path provided by the embodiment of the present specification;

[0054] Figure 2b is a schematic 3D effect diagram of the target puncture path provided by the embodiment of the present specification;

[0055] Figure 3 is a schematic diagram of the puncture constraint angle provided by the embodiment of the present specification;

[0056] Figure 4a is a schematic flowchart of determining the bile duct stone region and the bile duct tumor region provided by the embodiment of the present specification;

[0057] Figure 4b is a schematic diagram of a negative stone at low keV provided by the embodiment of the present specification;

[0058] Figure 4c is a schematic diagram of a negative stone at high keV provided by the embodiment of the present specification;

[0059] Figure 5a is a schematic flowchart of obtaining the liver multi-energy spectrum enhanced image sequence provided by the embodiment of the present specification;

[0060] Figure 5bSchematic diagram of the enhanced liver multi-energy spectrum image sequence provided by the embodiments of this specification;

[0061] Figure 6 Flow schematic diagram of obtaining the initial multi-energy spectrum image sequence provided by the embodiments of this specification;

[0062] Figure 7 Schematic diagram of the first target key area provided by the embodiments of this specification;

[0063] Figure 8 Schematic diagram of the second target key area provided by the embodiments of this specification;

[0064] Figure 9 Schematic diagram of the third target key area provided by the embodiments of this specification;

[0065] Figure 10 Schematic diagram of the xiphoid process area provided by the embodiments of this specification;

[0066] Figure 11 Schematic diagram of the right dilation penetration area provided by the embodiments of this specification;

[0067] Figure 12 Schematic diagram of the puncture target area provided by the embodiments of this specification;

[0068] Figure 13a Flow schematic diagram of determining the left puncture target area and the right puncture target area provided by the embodiments of this specification;

[0069] Figure 13b Schematic diagram of the bifurcation center point provided by the embodiments of this specification;

[0070] Figure 14 Schematic diagram of the puncture path generation device provided by the embodiments of this specification;

[0071] Figure 15 Internal structure diagram of the computer device provided by the embodiments of this specification. Detailed implementation manners

[0072] The embodiments of the present invention will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals represent the same or similar elements or elements with the same or similar functions from beginning to end. The embodiments described below by referring to the accompanying drawings are exemplary and are intended to explain the present invention, and should not be construed as a limitation to the present invention.

[0073] Percutaneous transhepatic cholangial drainage (PTCD) is a procedure that, under the guidance of X-ray or B-ultrasound, uses a special puncture needle to penetrate the intrahepatic bile duct percutaneously, and then directly injects contrast agent into the biliary tract to rapidly visualize the intrahepatic and extrahepatic bile ducts. At the same time, biliary tract drainage is performed through the contrast agent tube. PTCD is an important treatment method in biliary surgery and is widely used in the treatment of diseases such as cholangitis and cholangiocarcinoma.

[0074] The implementation of PTCD surgery requires precise preoperative planning, and the plain scan sequence of ordinary CT is difficult to extract all the key information required for the operation at one time, especially for the diagnosis of lesions such as negative stones or tumors. Compared with ordinary CT, multi-energy spectrum CT or photon CT can obtain images at multiple energies, improve the recognition accuracy of lesions such as negative stones and tumors, and reveal potential lesions such as stenosis, obstruction, negative stones and space-occupying tumors in the biliary tract. Therefore, multi-spectral CT or photon CT shows higher accuracy in the selection of the PTCD surgical path. PTCD surgery also faces many difficulties. For example, patients have poor physical conditions, low surgical tolerance and compliance; during the puncture process, it is difficult to locate the target bile duct, the puncture accuracy is poor, which may lead to vascular injury or mis-puncture into non-target areas, increasing the risk of bleeding and infection. In addition, the ultrasound guidance technology is limited by the experience and operation skills of the surgeon. These problems not only increase the postoperative complications and pain of patients, but also increase the medical costs, bringing a huge economic burden to patients and medical institutions.

[0075] The selection of the PTCD surgical puncture path needs to comprehensively consider various factors such as the patient's physical condition before surgery, the puncture risk during the operation, and the decompression effect after the operation. Finally, one or more puncture paths will be formulated based on these factors, and the key information of the operation will be clarified, including the skin puncture position, the needle insertion angle, the needle insertion depth, and the position of the puncture target, etc.

[0076] In related technologies, it mainly relies on determining the puncture angle and position, but usually uses traditional image processing methods for the segmentation of target tissues, mostly semi-automatic segmentation, and still requires a large amount of manual intervention. In the diagnosis and detection of bile duct stones, ERCP imaging technology is commonly used. These methods have limitations in surgical convenience and path generation accuracy, and cannot provide accurate diagnosis of negative stones or key information such as clear puncture paths and targets.

[0077] The solution for generating the PTCD surgical puncture path needs to include three key links:

[0078] 1. Recognition of stones and tumors in the bile duct by multi-energy spectrum CT:

[0079] There may be various types of stones, including positive stones (high-density stones) and negative stones (low-density stones). Positive stones are easily shown and identified under CT or X-ray, while negative stones such as cholesterol stones (cholesterol content above 70%), bilirubin stones (main component calcium bilirubinate), and mixed stones (containing cholesterol and calcium bilirubinate) are difficult to appear under conventional X-ray or CT images due to their low density.

[0080] In many patients undergoing PTCD surgery, tumors of the biliary tract system are common causes. Bile duct tumors can occur in the bile duct or common bile duct, and the early symptoms are usually not specific, so early diagnosis is challenging. Advanced bile duct tumors often lead to complete obstruction of the bile duct, which makes it crucial to quickly and accurately locate the tumor position for the smooth progress of the surgery.

[0081] In related technologies, endoscopic retrograde cholangiopancreatography (ERCP) is commonly used clinically for the diagnosis of intrahepatic bile duct stones. This method generates two-dimensional bile duct images through a dedicated C-shaped X-ray device. However, since it only provides planar images and cannot accurately locate the three-dimensional positions of stones or tumors, there are limitations in the preoperative planning. In addition, ERCP can only provide qualitative judgment and cannot be used for precise surgical planning.

[0082] In the screening of cholangiocarcinoma, enhanced CT and magnetic resonance imaging (MRI) are often applied. Although MRI can provide better image quality, its cost is high. Enhanced CT requires injection of iodine contrast agent, and some patients may have an allergic reaction to it. Therefore, multi-energy CT or photon CT becomes a more suitable choice, especially in the preoperative planning of PTCD surgery. Multi-energy CT uses X-rays with different energies (low kVp and high kVp) to improve the image quality of soft tissues and enhance the recognition accuracy of negative stones and intrahepatic bile duct tumors according to the different absorption characteristics of substances.

[0083] 2. Key region segmentation and key point extraction:

[0084] In medical images, the segmentation of key regions and the extraction of key points are important bases for accurate puncture path planning. Common key point extraction methods include traditional image processing techniques. However, the accuracy of traditional image processing methods is limited and the generalization ability is poor, and they cannot effectively adapt to the individual differences of different patients.

[0085] 3. Path constraints and path generation rules:

[0086] The generation of the puncture path needs to consider multiple constraints to ensure the safety and effectiveness of the surgery. The constraints of the path include: the distance between the puncture position and the rib, whether the path passes through important blood vessels in the liver (such as the hepatic artery, hepatic vein, and portal vein), whether there is a risk of repeated puncture of the bile duct, whether the angle between the puncture target and the bile duct course is acute, and whether the path passes through an intrahepatic space-occupying tumor, etc. If there are abnormalities such as negative stones, blood clots, infection foci, or tumors in the bile duct, it may cause blockage of the drainage tube, thereby affecting the postoperative drainage effect. In this case, the path generation rule should be capable of generating multiple puncture paths, and these paths should not be concentrated in the same biliary branch. This can effectively reduce the drainage problems caused by the blockage of a certain path and ensure the maximization of the postoperative decompression effect.

[0087] In the related technology, the threshold segmentation technology in traditional image processing is used to segment organ or tissue lesions. The specific steps include: First, search for data and find the point with the maximum gradient as the seed, and perform contour tracking to extract the set of points in the skin area. Then, use the seed region growing algorithm to obtain the puncture target area, and calculate the puncture angle and depth in the three-dimensional visualization scene to determine the puncture path.

[0088] However, the contour of the key area extracted by the threshold segmentation method is not clear. In addition, the generalization ability of the traditional image processing method is poor. When the image quality is poor and still relies on the traditional segmentation method, it may lead to incorrect area segmentation, which not only increases the workload of manual intervention but also may affect the segmentation accuracy. More importantly, this method does not consider the constraints on the puncture path and only relies on the puncture angle and depth to determine the puncture path, making it difficult to effectively reduce the risks during the surgery.

[0089] In the related technology, ERCP images are used as the neural network training dataset to identify bile ducts and positive stones. However, since ERCP images are two-dimensional images and cannot accurately reflect the specific positions in the three-dimensional space, they cannot provide precise support for the preoperative planning of PTCD surgery. In addition, this diagnosis and treatment plan has high requirements for the doctor's technology and experience, with a large operation difficulty, and also has relatively strict requirements for hospital equipment and medical teams. Therefore, the implementation of the plan has certain limitations and cannot effectively diagnose the existence of negative stones.

[0090] In the related technology, traditional image processing methods such as median filtering, edge detection, and Hough transform are used to extract the center points of key areas from the data, and the three-dimensional data is registered to achieve virtual reality applications, and then provide puncture navigation during the surgery. However, this method does not involve preoperative planning and only focuses on intraoperative guidance. And the generalization ability of traditional image processing methods is poor, and the center points cannot be accurately obtained.

[0091] In the related art, a traditional algorithm of region growing is used in energy spectrum CT images to extract the region of interest of stones; according to the region of interest of stones, a machine learning method is used for the composition analysis of stones. The specific method is to train the model using the random forest algorithm, and through multiple binary classification calculations, finally identify positive stone components such as calcium phosphate, calcium oxalate, and calcium oxalate dihydrate, and analyze them. However, this method has poor generalization ability and there will be cases of incorrect segmentation. In addition, the random forest algorithm is used in the composition analysis part to analyze positive stone components, but negative stones cannot be detected, which limits its application effect in the preoperative planning of PTCD surgery.

[0092] To sum up, the related art has the following main disadvantages:

[0093] First, there is a lack of effective identification of negative stones and intrahepatic bile duct tumors. Negative stones and intrahepatic bile duct tumors are crucial for PTCD surgical planning, so the detection of these lesions cannot be ignored during the operation.

[0094] Second, during the PTCD operation, the segmentation and extraction of key regions are not accurate enough. Relying on traditional image processing methods or having deficiencies in the segmentation and extraction of key regions results in low positioning accuracy.

[0095] Third, there is a lack of a complete and effective preoperative puncture path constraint method and generation rule, which cannot provide precise guidance and planning for the operation.

[0096] Based on this, the embodiments of this specification provide a puncture path generation method. First, a liver multi-energy spectrum enhanced image sequence including a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence is obtained. Then, the liver multi-energy spectrum enhanced image sequence is subjected to image segmentation to obtain a target key region including a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence. Then, based on the first target key region and the second target key region, the dilated penetration region and the puncture target region are determined. This process can reduce manual intervention, thereby improving the accuracy of determining the target puncture path. Finally, based on the dilated penetration region, the puncture target region, and the target key region, a target puncture path is generated. This embodiment determines the target puncture path based on the liver multi-energy spectrum enhanced image sequence, improves puncture safety and puncture accuracy, assists doctors in making more precise preoperative planning, and further provides more efficient and accurate surgical operation support for doctors, reducing the risks during the operation.

[0097] The embodiments of this specification provide a puncture path generation method. Please refer to Figure 1a , and this puncture path generation method may include the following steps:

[0098] S110. Obtain a liver multi-energy spectrum enhanced image sequence.

[0099] S120. Perform image segmentation on the liver multi-energy spectrum enhanced image sequence to obtain target key regions.

[0100] Among them, the liver multi-energy spectrum enhanced image includes a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence. The target key regions include a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence.

[0101] Specifically, use a medical imaging device to perform scans to obtain image sequences at different stages, and these image sequences respectively provide liver imaging information under different energy ranges and different blood flow dynamics in veins and arteries. After obtaining the image sequences, perform registration on them to ensure the spatial consistency between images in each phase, and obtain a liver multi-energy spectrum enhanced image sequence including a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence.

[0102] After completing the registration, use an image segmentation method (such as a deep learning-based model or a traditional segmentation method) to perform image segmentation on the liver multi-energy spectrum enhanced image sequence, identify and extract the first target key region from the multi-energy spectrum image sequence, identify and extract the second target key region from the venous phase image sequence, and identify and extract the third target key region from the arterial phase image sequence to obtain the target key regions. Exemplarily, combine the first target key region, the second target key region, and the third target key region into the venous phase image sequence to generate a comprehensive image containing all key regions. Please refer to Figure 1b , Figure 1b which shows the first target key region, the second target key region, and the third target key region of the venous phase image sequence.

[0103] S130. Determine the dilated penetration region and the puncture target region based on the first target key region and the second target key region.

[0104] S140. Generate a target puncture path based on the dilated penetration region, the puncture target region, and the target key regions.

[0105] Specifically, the generation process of the puncture path involves determining the penetration area and the puncture target area. First, by analyzing the first target key area and the second target key area, the area where puncture can be performed is determined, thereby determining the dilation penetration area and the puncture target area. Between the dilation penetration area and the puncture target area, penetration points and puncture points are selected in pixel units to generate multiple puncture paths. Finally, in combination with the target key area and the multiple puncture paths that have been generated, specific screening criteria are used to screen the multiple puncture paths, and the puncture paths that meet the requirements are selected as the target puncture paths.

[0106] It should be noted that the UI display and editing functions of the above-mentioned target key area and target puncture path include various forms of visual presentation methods, such as 2D images, 3D images, pseudo-color images, as well as curves, charts, etc. Users can perform interactive editing in these graphical interfaces. Images and data can be scaled, rotated, and the viewing angle can be switched according to needs to provide more comprehensive visual information. In addition, it also supports exporting the results in a printable format, generating a detailed report, and can be saved in multiple file formats for subsequent viewing and sharing.

[0107] In the above-mentioned embodiment, first, a liver multi-energy spectrum enhanced image sequence including a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence is obtained. Then, image segmentation is performed on the liver multi-energy spectrum enhanced image sequence to obtain a target key area including a first target key area associated with the multi-energy spectrum image sequence, a second target key area in the venous phase image sequence, and a third target key area in the arterial phase image sequence. Then, based on the first target key area and the second target key area, the dilation penetration area and the puncture target area are determined. This process can reduce manual intervention, thereby improving the accuracy of determining the target puncture path. Finally, based on the dilation penetration area, the puncture target area, and the target key area, the target puncture path is generated. This embodiment determines the target puncture path based on the liver multi-energy spectrum enhanced image sequence, improves puncture safety and puncture accuracy, assists the doctor in making a more accurate preoperative plan, and further provides more efficient and accurate surgical operation support for the doctor, reducing the risks during the operation.

[0108] In some embodiments, please refer to Figure 2a , generating the target puncture path based on the dilation penetration area, the puncture target area, and the target key area may include the following steps:

[0109] S210. Using the puncture path constraint conditions, based on the dilation penetration area, the puncture target area, and the target key area, determine the puncture path candidate set.

[0110] Specifically, the selection of the puncture path is restricted by various factors, such as the anatomical structure of the target area, the distribution of surrounding tissues, and the avoidance areas of important organs. Therefore, when planning the puncture path, these restrictions must be comprehensively considered to ensure that the selected path is both safe and effective. To accurately plan the puncture path, a series of puncture path constraint conditions need to be defined first to ensure that the path avoids key blood vessels, nerves, and other sensitive tissue areas. On this basis, between the dilation penetration area and the puncture target area, the penetration point and the puncture point are selected in pixels to generate multiple puncture paths. Next, using the puncture path constraint conditions combined with the target key area, the puncture paths that do not meet the conditions are filtered out from the multiple puncture paths, and the puncture paths that meet the puncture path constraint conditions are selected to form a puncture path candidate set.

[0111] S220. Use the distance calculation method to determine the target puncture path from the puncture path candidate set.

[0112] Specifically, the distance calculation method (such as Euclidean distance, Manhattan distance, etc.) is used to calculate the lengths of each puncture path in the puncture path candidate set. Then, by comparing the lengths of each puncture path in the puncture path candidate set, the shortest puncture path is selected from the puncture path candidate set as the target puncture path.

[0113] Exemplarily, calculate the Euclidean space distance of all puncture paths, and select the puncture path with the shortest distance as the target puncture path. Please refer to Figure 2b , and the penetration point (starting point), the puncture point (ending point), and the target puncture path (virtual puncture path) are fed back and presented in 3D in the form of a three-dimensional point set.

[0114] In the above embodiments, by using the puncture path constraint conditions, based on the dilation penetration area, the puncture target area, and the target key area, the puncture path candidate set is determined. Using the distance calculation method, the target puncture path with the shortest path and the least damage is determined from the puncture path candidate set, improving the puncture safety and puncture accuracy and reducing the risks during the operation.

[0115] In some embodiments, the puncture path constraint conditions may include: the puncture path does not pass through the hepatic artery area, the hepatic vein area, the portal vein area, and the rib area.

[0116] In some cases, the hepatic artery is one of the main blood vessels supplying blood to the liver and is responsible for bringing oxygenated blood to the liver. If the puncture path passes through the hepatic artery region, it may cause arterial bleeding or injury, leading to serious complications. The hepatic vein is responsible for carrying away the deoxygenated blood from the liver. If the puncture path passes through the hepatic vein region, it may cause venous injury or bleeding, affecting the blood return of the liver and potentially leading to impaired liver function. The portal vein is an important blood vessel responsible for transporting the blood from the intestines and spleen to the liver. If the portal vein region is damaged during the puncture process, it may result in bleeding or cause portal hypertension. The ribs not only protect important internal organs, but also if the puncture path passes through the rib region, it may cause the puncture needle to deflect or bend, increasing the difficulty of puncture or resulting in puncture failure. Therefore, the puncture path needs to avoid the above-mentioned regions.

[0117] Specifically, when planning the puncture path, it is necessary to calculate the intersections or overlaps between the puncture path and the hepatic artery region, hepatic vein region, portal vein region, and rib region. If the puncture path passes through any of the hepatic artery region, hepatic vein region, portal vein region, and rib region, then this puncture path is discarded. Then, all puncture paths are traversed, and the above calculation process is repeated to screen out one by one the puncture paths that do not pass through the hepatic artery, hepatic vein, portal vein, or rib region. These puncture paths will form a candidate set of puncture paths that meet the safety requirements. The set of puncture paths after screening will be used as the basis for subsequent selection and optimization of the target puncture path.

[0118] In addition, the presence of the ribs may limit the adjustment of the puncture needle angle, thus affecting the selection of the puncture path and further affecting the surgical effect. Therefore, it can be set that the Euclidean distance between the puncture path and the rib region should be greater than a preset range. If the distance is less than the preset range, then this path is discarded.

[0119] In the above embodiments, the puncture path does not pass through the hepatic artery region, hepatic vein region, portal vein region, and rib region, effectively improving the accuracy of puncture and reducing the surgical risk.

[0120] In some embodiments, the puncture path constraint condition may include: the puncture angle corresponding to the puncture path is less than a preset constraint angle.

[0121] Specifically, the selection of the preset constraint angle depends on multiple factors, including the patient's individual anatomical structure, the specific location and morphology of the bile duct, the technical requirements of the operation, etc. After establishing the preset constraint angle, the design of the puncture path needs to ensure that the puncture angle formed by the puncture path and the center line of the bile duct should be less than the preset constraint angle, so as to avoid difficulties in catheter insertion or operation failure caused by too large an angle, thus ensuring the smooth progress of the entire operation process. Therefore, if the puncture angle corresponding to the puncture path is not less than the preset constraint angle, then the puncture path is discarded. Traverse each puncture path and screen out one by one the puncture paths whose corresponding puncture angles are less than the preset constraint angle. These puncture paths will form a set of candidate puncture paths that meet the safety requirements. The set of puncture paths after screening will be used as the basis for subsequent selection and optimization of the target puncture path. Among them, the preset constraint angle can be set to 80 degrees.

[0122] Exemplarily, please refer to Figure 3 , Figure 3 the included angle between the virtual puncture path and the center line of the bile duct (bile duct running line) in

[0123] In the above embodiments, the puncture angle corresponding to the puncture path is less than the preset constraint angle, effectively improving the accuracy of puncture and reducing the surgical risk.

[0124] In some embodiments, the puncture path constraint conditions may include: the puncture path does not repeatedly pass through the same branch bile duct.

[0125] Specifically, the puncture path is constrained to avoid repeatedly passing through the same branch bile duct and avoid puncturing the bile duct that has already been punctured, because repeated puncture may increase the risk of biliary tract injury and thus cause complications, such as bile fistula (bile leakage) or biliary tract bleeding. Therefore, if the puncture path repeatedly passes through the same branch bile duct, then the puncture path is discarded. Traverse each puncture path and screen out one by one the puncture paths that do not repeatedly pass through the same branch bile duct. These puncture paths will form a set of candidate puncture paths that meet the safety requirements. The set of puncture paths after screening will be used as the basis for subsequent selection and optimization of the target puncture path.

[0126] In the above embodiments, the puncture path does not repeatedly pass through the same branch bile duct, effectively improving the accuracy of puncture and reducing the surgical risk.

[0127] In some embodiments, the puncture path constraint conditions may include: the puncture path does not pass through bile duct stones and bile duct tumors.

[0128] Specifically, if the puncture path traverses the bile duct stones, it may cause the displacement of the stones, thus making the condition more complex; if the puncture path involves a bile duct tumor, the puncture needle may damage the tumor tissue, leading to the implantation and metastasis of tumor cells, and further aggravating the condition. Therefore, when selecting the puncture path, the bile duct stones and tumors should be avoided. So, if there are bile duct stones and bile duct tumors in the bile duct where the puncture path is located, the puncture path should be discarded. Traverse each puncture path, and screen out the puncture paths that do not pass through the bile duct stones and bile duct tumors one by one. These puncture paths will form a candidate set of puncture paths that meet the safety requirements. The screened puncture path set will be used as the basis for the subsequent selection and optimization of the target puncture path.

[0129] In the above embodiments, the puncture path does not pass through the bile duct stones and bile duct tumors, effectively improving the accuracy of the puncture and reducing the surgical risk.

[0130] In some embodiments, please refer to Figure 4a , the multi-energy spectrum image sequence includes a plain scan image sequence and an energy spectrum image sequence, the bile duct stones include positive stones and negative stones, and the method further includes determining the bile duct stone area and the bile duct tumor area by the following methods:

[0131] S410. Segment the bile duct area in the plain scan image sequence using a preset density to determine the positive stone area.

[0132] Specifically, high keV (kilo-electron volt) images can present the different density and structural characteristics of tissues more clearly due to their strong penetration ability and low noise level. Therefore, in the multi-energy spectrum image sequence, select the high keV images from the multi-energy spectrum image sequence as the plain scan image sequence. The low keV images are used as the energy spectrum image sequence, which is suitable for better evaluating the element distribution and composition differences in tissues.

[0133] The pixels of the image are divided into different regions, and each region corresponds to different tissue or structural characteristics. Therefore, according to a large amount of experimental data and clinical experience of the regions to be segmented as needed, preset the preset density. Use the threshold segmentation technology to process the plain scan image sequence, and segment the regions in the bile duct area that meet the preset density, and mark them as the positive stone area.

[0134] S420. Determine the energy spectrum curve based on the bile duct area in the multi-energy spectrum image sequence.

[0135] S430. Determine the negative stone and bile duct tumor areas based on the slope of the energy spectrum curve.

[0136] Specifically, based on its density manifestation and imaging law, the spectral curves of the bile duct region in the multi-energy spectral image sequence can be calculated. The multi-energy spectral image sequence includes a plain scan image sequence with high keV and an energy spectral image sequence with low keV. Then, by fitting the slopes of the spectral curves of these two energy spectral image sequences, the imaging characteristics at different energies can be obtained. Since the compositions of negative stones and bile duct tumors are similar to fat, their attenuation rates are higher in the low keV region. By calculating the difference between the slope of the spectral curve corresponding to the plain scan image sequence and the slope of the spectral curve corresponding to the energy spectral image sequence, it is possible to identify whether there is a region with a negative slope. If a negative value region appears, it can be determined that this region is a negative stone or bile duct tumor region, but at this time, negative stones and bile duct tumors are not distinguished. Exemplarily, please refer to Figure 4b , Figure 4b which shows that negative stones present low density at low keV (40 keV). Please refer to Figure 4c , Figure 4c which shows that negative stones present high density at high keV (140 keV).

[0137] It should be noted that in order to evaluate the risk of bile duct obstruction caused by stones, it is necessary to combine the proportion of positive stones or bile duct tumors and negative stones within the cross-sectional area of the bile duct, and then judge whether the stones will cause biliary obstruction or lead to biliary stenosis. This evaluation result can be used as a trigger condition for multiple puncture paths, and multiple puncture paths can be generated based on the risk assessment. It should be noted that the generated puncture paths should avoid being within the same biliary branch to ensure accurate coverage of potential lesion areas.

[0138] In the above embodiments, the bile duct region in the plain scan image sequence is segmented using a preset density to determine the positive stone region. Based on the bile duct region in the multi-energy spectral image sequence, the spectral curve is determined. Based on the slope of the spectral curve, the negative stone and bile duct tumor regions are determined, providing a basis for subsequent determination of a more accurate target puncture path to avoid the puncture path passing through stones or tumors.

[0139] In some embodiments, please refer to Figure 5a , the steps for obtaining the multi-energy spectral enhanced image sequence of the liver may include the following:

[0140] S510. Obtain the initial multi-energy spectral enhanced image sequence of the liver.

[0141] S520. Based on the venous phase image sequence, perform rigid registration on the initial multi-energy spectral image sequence and the initial arterial phase image sequence to obtain the multi-energy spectral enhanced image sequence of the liver.

[0142] Among them, the initial multi-energy spectral enhanced image sequence of the liver includes the venous phase image sequence, the initial multi-energy spectral image sequence, and the initial arterial phase image sequence.

[0143] Specifically, first, use a medical imaging device to perform scans at different energy levels to obtain an initial multi-energy spectrum image sequence. For example, multi-energy spectrum computed tomography (CT) technology can be adopted, and scans are performed using multiple X-rays with different energies to obtain an initial multi-energy spectrum image sequence. When X-rays with different energies penetrate tissues, their attenuation degree is closely related to the atomic number of the substance and the ray energy. Therefore, different attenuation information can be provided, and the unique attenuation curves of different substances can be captured. Next, inject a contrast agent and perform a scan during the arterial phase to reflect the blood flow in the hepatic artery and obtain an initial arterial phase image sequence. Then, continue the scan to obtain a venous phase image sequence for observing the imaging changes during the venous phase. Select the venous phase image sequence as the reference image for registration with other image sequences. Based on the venous phase image sequence, use rigid transformation (including operations such as translation and rotation) to register the initial multi-energy spectrum image sequence and the initial arterial phase image sequence, so that images at different time points or different energy levels are spatially aligned, ensuring accurate matching of pixel positions between different images, thereby obtaining a hepatic multi-energy spectrum enhanced image sequence. The hepatic multi-energy spectrum enhanced image sequence integrates image information at different energy levels and ensures precise alignment between images in each phase, providing reliable image data for subsequent determination of the puncture path.

[0144] Exemplarily, please refer to Figure 5b , Figure 5b where 502 in Figure 5b shows the registered multi-energy spectrum image sequence, Figure 5b where 504 in

[0145] shows the registered arterial phase image sequence, and

[0146] where 506 in Figure 6 shows the venous phase image sequence.

[0147] S610. Set multiple energy thresholds.

[0148] S620. Collect photon counting data at each energy threshold and perform image reconstruction to generate an initial multi-energy spectrum image sequence.

[0149] Specifically, different X-ray energies can affect the absorption mode of substances, thereby affecting the imaging effect of images. Therefore, in practical applications, multiple energy thresholds can be set according to requirements to obtain various information about the composition and structure of substances. Photon CT can directly count X-ray photons and classify them according to their energies, and can collect photon count data at each energy threshold in a single scan. Using an image reconstruction algorithm to reconstruct the photon count data at each energy threshold to generate an image corresponding to each energy threshold, and then combining the images at each energy threshold to obtain an initial multi-energy spectrum image sequence. Since the initial multi-energy spectrum image sequence can capture the unique attenuation curves of different substances, it can improve the accuracy of subsequent identification of negative stones and tumors, thereby improving the accuracy of puncture path selection.

[0150] In the above embodiment, multiple energy thresholds are set, the photon count data at each energy threshold is collected and image reconstruction is performed to generate an initial multi-energy spectrum image sequence, providing a data basis for subsequent identification of negative stones and tumors.

[0151] In some embodiments, the multi-energy spectrum image sequence includes a plain scan image sequence. Image segmentation is performed on the liver multi-energy spectrum enhanced image sequence to obtain a target key region, which may include: using a plain scan segmentation model to perform image segmentation on the plain scan image sequence to obtain a first target key region.

[0152] Specifically, due to its strong penetration ability and low noise level, the high keV (kilo-electron volt) image can more clearly present the different density and structural characteristics of tissues. Therefore, in the multi-energy spectrum image sequence, the high keV image is selected from the multi-energy spectrum image sequence as the plain scan image sequence. Next, the plain scan image sequence is used as the input of the plain scan segmentation model, and the plain scan segmentation model will perform image segmentation on each frame of the plain scan image sequence, identify and distinguish different target regions in the image, thereby obtaining the first target key region. Among them, the first target key region includes the xiphoid process region, rib region, liver region and skin contour region.

[0153] Exemplarily, since the xiphoid process region, rib region, liver region and skin contour region have obvious differences in spatial distribution and texture features, these regions can be effectively segmented by machine learning techniques. Based on a deep learning network model of machine learning, an annotated plain scan image sequence sample is input, where the annotation includes the xiphoid process region, rib region, liver region and skin contour region. Through data preprocessing, model optimization and deep learning training, a trained plain scan segmentation model is obtained.

[0154] In the actual application process, the plain scan image sequence is input into the trained plain scan segmentation model, and the xiphoid process region, rib region, liver region and skin contour region are output. Please refer to Figure 7 ,Figure 7 The red area in Figure 7 the yellow area represents the skin contour area, Figure 7 the green area represents the rib area, Figure 7 and the blue area represents the xiphoid process area.

[0155] In the above embodiments, the plain scan segmentation model is used to segment the plain scan image sequence to obtain the first target key area, providing a data basis for subsequent determination of the target puncture path.

[0156] In some embodiments, segmenting the liver multi-energy spectrum enhanced image sequence to obtain the target key area may include: using the vein segmentation model to segment the venous phase image sequence to obtain the second target key area.

[0157] Specifically, taking the venous phase image sequence as the input of the vein segmentation model, the vein segmentation model will segment each frame of the venous phase image sequence, identify and distinguish different target areas in the image, so as to obtain the second target key area. Among them, the second target key area includes the portal vein area, hepatic vein area, gallbladder area, and bile duct area.

[0158] Exemplarily, since the portal vein area, hepatic vein area, gallbladder area, and bile duct area have obvious differences in spatial distribution and texture features, these areas can be effectively segmented by machine learning techniques. Based on the deep learning network model of machine learning, input the labeled venous phase image sequence samples, where the labels include the portal vein area, hepatic vein area, gallbladder area, and bile duct area. Through data preprocessing, model optimization, and deep learning training, a trained vein segmentation model is obtained.

[0159] In the actual application process, input the venous phase image sequence into the trained vein segmentation model, and output the portal vein area, hepatic vein area, gallbladder area, and bile duct area. Please refer to Figure 8 , Figure 8 the blue area in Figure 8 the yellow area represents the hepatic vein area, Figure 8 the green area represents the gallbladder area, Figure 8 and the red area represents the bile duct area.

[0160] In the above embodiments, the vein segmentation model is used to segment the venous phase image sequence to obtain the second target key area, providing a data basis for subsequent determination of the target puncture path.

[0161] In some embodiments, image segmentation of the liver multi-energy spectrum enhanced image sequence to obtain the target key regions may include: using an artery segmentation model to perform image segmentation on the arterial phase image sequence to obtain the third target key region.

[0162] Specifically, taking the arterial phase image sequence as the input of the artery segmentation model, the artery segmentation model will perform image segmentation on each frame of the arterial phase image sequence, identify and distinguish different target regions in the image, so as to obtain the third target key region. Among them, the third target key region includes the hepatic artery region and intrahepatic tumors.

[0163] Exemplarily, since the hepatic artery region and intrahepatic tumors have obvious differences in spatial distribution and texture features, these regions can be effectively segmented by machine learning techniques. Based on a deep learning network model of machine learning, an annotated arterial phase image sequence sample is input, where the annotation includes the hepatic artery region and intrahepatic tumors. Through data preprocessing, model optimization, and deep learning training, a trained artery segmentation model is obtained.

[0164] In the actual application process, the arterial phase image sequence is input into the trained artery segmentation model, and the hepatic artery region and intrahepatic tumors are output. Please refer to Figure 9 , Figure 9 where the green region in Figure 9 represents the hepatic artery region, Figure 9 the blue region in

[0165] represents intrahepatic tumors, and

[0166] the red region in

[0167] represents the abdominal aorta.

[0168]

[0169] Among them, W is the weight factor, X represents the feature maps corresponding to different layers of decoders, and Y represents the feature maps after multi-scale supervision.

[0170] For the key regional information of blood vessels, a structural similarity (SSIM) loss function is introduced. The calculation formula of this loss function is as follows:

[0171]

[0172] Among them, μ X is the average value of X, is the variance of X, σ XY is the covariance of XY, and c1, c2 are constants to maintain stability and can be obtained through the following formula:

[0173] c1 = (k1L) 2

[0174] c2 = (k2L) 2

[0175] Among them, k1 and k2 are 0.01 and 0.03 respectively, and L is the dynamic range of pixel values.

[0176] In some embodiments, the first target key region includes the xiphoid process region, the rib region, and the liver region, and the dilation penetration region includes the left dilation penetration region and the right dilation penetration region. Based on the first target key region and the second target key region, determining the dilation penetration region and the puncture target region may include: mapping based on the xiphoid process region to determine the left dilation penetration region.

[0177] Specifically, the determination of the puncture path can be selected based on the left or right region of the human body. When the left side is selected as the puncture direction, the multi-energy spectrum enhanced image sequence of the liver after image segmentation is adjusted to the coronal plane. Subsequently, the xiphoid process region is mapped onto the chest skin contour. The key to this mapping process is to accurately locate the corresponding point of the xiphoid process on the skin surface to ensure the correctness and safety of the puncture path. Next, based on the mapped region, the region for puncture entry is delimited within the T range, denoted as the left dilation penetration region. Among them, the T range is usually defined as 1 to 2 centimeters according to clinical research. Exemplarily, please refer to Figure 10 , Figure 10 The position indicated by the arrow in

[0178] is the xiphoid process region.

[0179] In some embodiments, the first target key region includes the xiphoid region, the rib region, and the liver region. The dilation penetration regions include a left dilation penetration region and a right dilation penetration region. Based on the first target key region and the second target key region, determining the dilation penetration region and the puncture target region may include: mapping based on the rib region and the liver region to determine the right dilation penetration region.

[0180] Specifically, the determination of the puncture path can be selected based on the left or right region of the human body. When the right side is selected as the puncture direction, it is first necessary to identify and determine the rib region, especially the specific positions and names of the 8th to 10th ribs. Adjust the multi-energy spectrum enhanced image sequence of the segmented liver to the coronal plane. Next, map the 8th to 10th ribs to the right skin contour. According to this mapping result, define the upper boundary of the puncture region as the upper edge of the 8th rib and the lower boundary as the lower edge of the 10th rib. Then, on the right skin contour, draw the midaxillary line perpendicular to the chest wall through the center of the armpit. Project the liver region onto the right skin contour, and the boundaries of the mapped region of the liver region are centered on the midaxillary line to define the anterior chest boundary and the back boundary of the puncture region. The upper boundary, lower boundary, anterior chest boundary, and back boundary of the puncture region form the right dilation penetration region. Exemplarily, please refer to Figure 11 , Figure 11 the red region in is the right dilation penetration region.

[0181] In the above embodiments, mapping based on the rib region and the liver region to determine the right dilation penetration region provides a data basis for subsequent determination of the target puncture path.

[0182] In some embodiments, the first target key region includes the xiphoid region, the rib region, and the liver region. The dilation penetration regions include a left dilation penetration region and a right dilation penetration region. Based on the first target key region and the second target key region, determining the dilation penetration region and the puncture target region may include: screening and partitioning based on the second target key region to determine the puncture target region.

[0183] Specifically, after determining the penetration region, it is also necessary to determine the puncture region to form the puncture path. Therefore, according to the screening and partitioning conditions, determine the key region that meets the requirements from the second target key region and determine it as the puncture target region. Exemplarily, please refer to Figure 12 , Figure 12 the red region in is the puncture target region.

[0184] In the above embodiments, screening and partitioning based on the second target key region to determine the puncture target region provides a data basis for subsequent determination of the target puncture path.

[0185] In some embodiments, please refer to Figure 13a, the second target key area includes the bile duct area. Based on the second target key area, screening and division are performed to determine the puncture target area, which may include the following steps:

[0186] S1310. Extract the centerline corresponding to the bile duct area.

[0187] Among them, the centerline corresponds to a bifurcation center point.

[0188] Specifically, use a centerline extraction algorithm (such as distance transformation, curve fitting) to extract the bile duct area to determine the centerline corresponding to the bile duct area. Then, analyze the running trend and bifurcation position of the centerline to identify the changes in its direction and the bifurcation position. In particular, for the bifurcation points of the left and right branches of the bile duct, use an algorithm to calculate the coordinates of the bifurcation points and mark them as bifurcation center points.

[0189] Exemplarily, please refer to Figure 13b , Figure 13b in which the pink area represents the liver area, Figure 13b in which the green area represents the gallbladder area, Figure 13b in which the red area represents the bile duct area, Figure 13b and point A in it represents the bifurcation center point between the left hepatic duct and the right hepatic duct in the bile duct area.

[0190] S1320. Determine a set of puncture target candidates from the bile duct area based on preset screening conditions and the centerline.

[0191] S1330. Divide the set of puncture target candidates into a left puncture target area and a right puncture target area based on the bifurcation center point.

[0192] Specifically, first calculate by walking along the centerline of the bile duct to determine the diameter of the branch bile duct at each position. Then, compare the diameter of each branch bile duct with the preset screening criteria, and screen out the qualified branch bile ducts from them to form a set of puncture target candidates. Next, for the bile duct areas in the set of puncture target candidates, evaluate them according to their spatial relationship with the bifurcation center point between the left and right hepatic ducts. By analyzing the spatial positions of these bile ducts, further divide the set of puncture target candidates into a left puncture target area and a right puncture target area.

[0193] Exemplarily, in clinical research, it is found that the more dilated branch bile ducts in the liver are easier to perform puncture guidance. Therefore, the screening condition can be set as the branch bile ducts dilated to grade 3 to 4.

[0194] In the above embodiments, the central line corresponding to the bile duct region is extracted, and a set of candidate puncture targets is determined from the bile duct region based on preset screening conditions and the central line. The set of candidate puncture targets is divided into a left puncture target region and a right puncture target region based on the bifurcation center point, which improves the success rate and safety of the puncture operation and provides a data basis for determining the target puncture path subsequently.

[0195] Embodiments of this specification provide a puncture path generation device 1400. Please refer to Figure 14 , the puncture path generation device 1400 includes: a multi-energy spectrum enhanced image acquisition module 1410, a target key region segmentation module 1420, a penetration and target region determination module 1430, and a target puncture path generation module 1440.

[0196] The multi-energy spectrum enhanced image acquisition module 1410 is configured to acquire a sequence of multi-energy spectrum enhanced liver images, wherein the multi-energy spectrum enhanced liver images include a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence;

[0197] The target key region segmentation module 1420 is configured to perform image segmentation on the sequence of multi-energy spectrum enhanced liver images to obtain a target key region, wherein the target key region includes a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence;

[0198] The penetration and target region determination module 1430 is configured to determine a dilation penetration region and a puncture target region based on the first target key region and the second target key region;

[0199] The target puncture path generation module 1440 is configured to generate a target puncture path based on the dilation penetration region, the puncture target region, and the target key region.

[0200] For the specific description of the puncture path generation device, reference may be made to the description of the puncture path generation method in the foregoing text, which will not be elaborated herein.

[0201] Embodiments of this specification provide a medical imaging device, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the method steps in the above embodiments are implemented.

[0202] In some embodiments, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as Figure 15As shown in the figure. The computer device includes a processor, a memory, a communication interface, a display screen, and an input device connected through a system bus. Among them, 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 the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a carrier network, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a puncture path generation method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, touchpad, or mouse, etc.

[0203] Those skilled in the art can understand that Figure 15 the structure shown in the figure is only a block diagram of some parts of the structure related to the solution disclosed in this specification, and does not constitute a limitation on the computer device to which the solution disclosed in this specification is applied. Specifically, the computer device may include more or fewer components than those shown in the figure, or combine some components, or have a different component arrangement.

[0204] In some embodiments, a computer device is provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the method steps in the above embodiments are implemented.

[0205] An embodiment of this specification provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in any of the above embodiments are implemented.

[0206] An embodiment of this specification provides a computer program product. The computer program product includes instructions. When the instructions are executed by the processor of the computer device, the computer device can execute the steps of the method in any of the above embodiments.

[0207] Note that the logic and / or steps represented in the flowchart or described otherwise herein, for example, can be considered as a definite sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection part having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable media on which a program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or other appropriate processing as necessary, and then stored in a computer memory.

Claims

1. A puncture path generation method, characterized in that The method includes: Obtaining a liver multi-energy spectrum enhanced image sequence, where the liver multi-energy spectrum enhanced image includes a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence; Performing image segmentation on the liver multi-energy spectrum enhanced image sequence to obtain a target key region, where the target key region includes a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence; Based on the first target key region and the second target key region, determining an expanded penetration region and a puncture target region; Generating a target puncture path based on the expanded penetration region, the puncture target region, and the target key region.

2. The method according to claim 1, wherein The generating a target puncture path based on the expanded penetration region, the puncture target region, and the target key region includes: Using puncture path constraint conditions, based on the expanded penetration region, the puncture target region, and the target key region, determining a set of puncture path candidates; Using a distance calculation method to determine a target puncture path from the set of puncture path candidates.

3. The method according to claim 2, characterized in that, The puncture path constraint conditions include: The puncture path does not pass through the hepatic artery region, the hepatic vein region, the portal vein region, or the rib region; The puncture angle corresponding to the puncture path is less than a preset constraint angle; The puncture path does not repeatedly pass through the same branch bile duct; The puncture path does not pass through bile duct stones and bile duct tumors.

4. The method according to claim 3, wherein The multi-energy spectrum image sequence includes a plain scan image sequence and an energy spectrum image sequence, the bile duct stones include positive stones and negative stones, and the method further includes determining the bile duct stone region and the bile duct tumor region by the following methods: Using a preset density to segment the bile duct region in the plain scan image sequence to determine the positive stone region; Based on the bile duct region in the multi-energy spectrum image sequence, determining an energy spectrum curve; Based on the slope of the energy spectrum curve, determining the negative stone and the bile duct tumor region.

5. The method according to claim 1, characterized in that The obtaining a liver multi-energy spectrum enhanced image sequence includes: Obtaining an initial liver multi-energy spectrum enhanced image sequence, where the initial liver multi-energy spectrum enhanced image sequence includes the venous phase image sequence, an initial multi-energy spectrum image sequence, and an initial arterial phase image sequence; Based on the venous phase image sequence, performing rigid registration on the initial multi-energy spectrum image sequence and the initial arterial phase image sequence to obtain the liver multi-energy spectrum enhanced image sequence.

6. The method according to claim 5, characterized in that, Obtaining the initial multi-energy spectrum image sequence by photon CT includes: Setting multiple energy thresholds; Collecting photon count data at each energy threshold and performing image reconstruction to generate the initial multi-energy spectrum image sequence.

7. The method according to claim 1, wherein The multi-energy spectrum image sequence includes a plain scan image sequence, and the performing image segmentation on the liver multi-energy spectrum enhanced image sequence to obtain a target key region includes: Using a plain scan segmentation model to perform image segmentation on the plain scan image sequence to obtain the first target key region; Using a venous segmentation model to perform image segmentation on the venous phase image sequence to obtain the second target key region; Use the arterial segmentation model to perform image segmentation on the arterial phase image sequence to obtain the third target key region.

8. The method according to claim 1, wherein The first target key region includes the xiphoid process region, the rib region, and the liver region. The penetration region includes a left penetration region and a right penetration region. Determining the penetration region and the puncture target region based on the first target key region and the second target key region includes: Based on mapping of the xiphoid process region, determine the left penetration region; Based on mapping of the rib region and the liver region, determine the right penetration region; Based on screening and partitioning of the second target key region, determine the puncture target region.

9. The method according to claim 8, characterized in that, The second target key region includes the bile duct region. Determining the puncture target region based on screening and partitioning of the second target key region includes: Extract the centerline corresponding to the bile duct region, where the centerline corresponds to a bifurcation center point; Based on a preset screening condition and the centerline, determine a puncture target candidate set from the bile duct region; Based on the bifurcation center point, divide the puncture target candidate set into a left puncture target region and a right puncture target region.

10. A puncture path generation device, characterized in that, The device includes: A multi-energy spectrum enhanced image acquisition module for acquiring a liver multi-energy spectrum enhanced image sequence, where the liver multi-energy spectrum enhanced image includes a multi-energy spectrum image sequence, a venous phase image sequence, and an arterial phase image sequence; A target key region segmentation module for performing image segmentation on the liver multi-energy spectrum enhanced image sequence to obtain target key regions, where the target key regions include a first target key region associated with the multi-energy spectrum image sequence, a second target key region in the venous phase image sequence, and a third target key region in the arterial phase image sequence; A penetration and target region determination module for determining a penetration region and a puncture target region based on the first target key region and the second target key region; A target puncture path generation module for generating a target puncture path based on the penetration region, the puncture target region, and the target key region.

11. A medical imaging device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

12. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

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

Images