Image registration method and device, computer equipment, storage medium and program product
During the image registration process, rigid adjustment is performed based on the elastic registration displacement field of the tubular structure mask, the problem of vascular morphological distortion is solved and the accuracy of image registration is improved.
Patent Information
- Application Number
- CN202311007848.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-08-09
- Publication Date
- 2025-02-18
AI Technical Summary
The prior art can easily lead to vascular morphological distortion when performing image registration, increasing the risk of vascular interspersion.
By obtaining the initial registration displacement field between the reference image of the target site and the floating image, the elastic registration displacement field is obtained based on the tubular structure mask, and rigid adjustment is performed in the overlapping area to obtain the rigid displacement field of the overlapping area, and finally based on this The image registration results were obtained from the two displacement fields.
It effectively avoids vascular morphological distortion, improves the accuracy of image registration, and reduces the risk of vascular interspersion.
Smart Images
Figure CN119477998A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of image processing technology, and in particular to an image registration method, apparatus, computer equipment, storage medium and program product. Background Art
[0002] With the continuous development of image registration technology, image registration is increasingly used in the medical field. Taking the image registration of the liver at different stages as an example, due to the asymmetry of vascular information at different stages, it will lead to abnormalities in the anatomical position and structure of the blood vessels, resulting in the risk of vascular interpenetration.
[0003] In order to reduce the risk of vascular interpenetration, the related technology mainly finds the possible interpenetrating blood vessels from the blood vessels in the images of different phases based on the position information of the blood vessels, and processes the possible interpenetrating blood vessels to reduce the vascular interpenetration phenomenon in the image registration process.
[0004] However, the related art methods may cause the problem of vascular morphology distortion. Summary of the invention
[0005] Based on this, it is necessary to provide an image registration method, device, computer equipment, storage medium and program product to solve the problem of vascular morphological distortion in response to the above technical problems.
[0006] In a first aspect, the present application provides an image registration method, the method comprising:
[0007] obtaining an initial registration displacement field between the reference image and the floating image of the target part;
[0008] Based on the initial registration displacement field, an elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part is obtained;
[0009] Based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image, the elastic registration displacement field is rigidly adjusted to obtain the rigid displacement field of the overlapping area;
[0010] Based on the rigid displacement field of the overlapping area and the initial registration displacement field, the image registration result between the reference image and the floating image is obtained.
[0011] In one embodiment, based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image, the elastic registration displacement field is rigidly adjusted to obtain the rigid displacement field of the overlapping area, including:
[0012] Determine the elastic registration displacement field to be adjusted corresponding to the overlapping area from the elastic registration displacement field;
[0013] Performing transformation estimation on the elastic registration displacement field to be adjusted to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted;
[0014] The estimated transformation model is used to perform rigid adjustment on the elastic registration displacement field to be adjusted, and the rigid displacement field of the overlapping area is obtained.
[0015] In one embodiment, performing transformation estimation on the elastic registration displacement field to be adjusted to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted includes:
[0016] Dividing the elastic registration displacement field to be adjusted to obtain a plurality of local elastic registration displacement fields to be adjusted;
[0017] Selecting a plurality of initial displacement vectors from each local elastic registration displacement field to be adjusted to obtain an initial estimated transformation model;
[0018] Multiple iterations are performed on each initial estimated transformation model until each initial estimated transformation model meets a preset condition, thereby obtaining an estimated transformation model corresponding to the elastic registration displacement field to be adjusted.
[0019] In one embodiment, multiple iterations are performed on each initial estimated transformation model, including:
[0020] For any iteration process of any initial estimated transformation model, obtaining the projection errors between all displacement vectors in the local elastic registration displacement field to be adjusted and the estimated transformation model obtained in the previous iteration;
[0021] Taking the displacement vector whose projection error is less than the preset error threshold as the iterative displacement vector;
[0022] Based on the iterative displacement vector, the estimated transformation model obtained in the previous iteration is iterated.
[0023] In one embodiment, the elastic registration displacement field to be adjusted is rigidly adjusted using the estimated transformation model to obtain the rigid displacement field of the overlapping area, including:
[0024] Using the estimated transformation model, a local rigid transformation is performed on the elastic registration displacement field to be adjusted to obtain multiple local rigid displacement fields;
[0025] The local rigid displacement fields are fused to obtain the rigid displacement field in the overlapping area.
[0026] In one embodiment, obtaining an elastic registration displacement field between a reference image tubular structure mask and a floating image tubular structure mask of a target part based on an initial registration displacement field includes:
[0027] The floating image tubular structure mask is sampled and transformed using the initial registration displacement field;
[0028] Performing color inversion processing on the tubular structure mask of the floating image after sampling transformation and the tubular structure mask of the reference image;
[0029] The tubular structure mask of the floating image after inversion processing and the tubular structure mask of the reference image are elastically registered to obtain an elastic registration displacement field.
[0030] In one embodiment, obtaining an image registration result between a reference image and a floating image based on a rigid displacement field of an overlapping area and an initial registration displacement field includes:
[0031] The rigid displacement field of the overlapping area is combined with the initial registration displacement field to obtain a composite displacement field between the reference image and the floating image;
[0032] Based on the composite displacement field, the image registration result between the reference image and the floating image is obtained.
[0033] In one embodiment, the method further comprises:
[0034] Obtain a mask heat map corresponding to the tubular structure mask of the floating image; different probability ranges in the mask heat map correspond to different identification information;
[0035] Interpolate and sample the mask heat map based on the composite displacement field;
[0036] According to the mask heat map after interpolation sampling and the preset probability threshold, the image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image is obtained.
[0037] In a second aspect, the present application further provides an image registration device, the device comprising:
[0038] A first acquisition module is used to acquire an initial registration displacement field between a reference image and a floating image of a target part;
[0039] A second acquisition module is used to acquire an elastic registration displacement field between a reference image tubular structure mask and a floating image tubular structure mask of a target part based on the initial registration displacement field;
[0040] An adjustment module, used for rigidly adjusting the elastic registration displacement field based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image to obtain a rigid displacement field of the overlapping area;
[0041] The determination module is used to obtain the image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field.
[0042] In a third aspect, the present application further provides a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the content of any one of the above-mentioned image registration method embodiments in the first aspect is implemented.
[0043] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the content of any one of the above-mentioned image registration method embodiments in the first aspect is implemented.
[0044] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the content of any one of the above-mentioned image registration method embodiments in the first aspect when executed by a processor.
[0045] The above-mentioned image registration method, device, computer equipment, storage medium and program product obtain the initial registration displacement field between the reference image and the floating image of the target part; based on the initial registration displacement field, obtain the elastic registration displacement field between the tubular structure mask of the reference image and the tubular structure mask of the floating image of the target part; based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image, rigidly adjust the elastic registration displacement field to obtain the rigid displacement field of the overlapping area; based on the rigid displacement field of the overlapping area and the initial registration displacement field, obtain the image registration result of the reference image and the floating image. The method uses the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image to rigidly adjust the elastic registration displacement field between the tubular structure mask of the reference image and the tubular structure mask of the floating image, so that the overlapping tubular structures can be separated and processed, thereby avoiding the morphological distortion of the tubular structures in the registration process due to the overlap. Based on the adjusted rigid displacement field of the overlapping area and the initial registration displacement field, the problem of morphological distortion of the tubular structure in the image registration results between the reference image and the floating image is avoided, thereby improving the accuracy of image registration. BRIEF DESCRIPTION OF THE DRAWINGS
[0046] Figure 1 A diagram of an application environment of an image registration method in an embodiment;
[0047] Figure 2 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0048] Figure 3 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0049] Figure 4 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0050] Figure 5is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0051] Figure 6 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0052] Figure 7 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0053] Figure 8 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0054] Fig. 9 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0055] Fig.10 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0056] Fig.11 is a schematic diagram of a flow chart of an image registration method in one embodiment;
[0057] Fig.12 is a schematic diagram of a blood vessel image registration result in one embodiment;
[0058] Fig.13 is a schematic diagram of a blood vessel image registration result in one embodiment;
[0059] Fig.14 is a schematic diagram of a blood vessel image registration result in one embodiment;
[0060] Fig.15 is a schematic diagram of a blood vessel image registration result in one embodiment;
[0061] Fig.16 FIG. 4 is a structural block diagram of an image registration device in one embodiment. DETAILED DESCRIPTION
[0062] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.
[0063] Before introducing the technical solution of the present application in detail, a brief introduction to the background technology of the present application is first given.
[0064] Taking the image registration of the liver at different phases as an example, the multi-phase enhanced examination of the liver is to scan at different delayed time points after intravenous injection of contrast agent to obtain the internal structure of different tissues and organs and lesions in the liver. Among them, the multi-phase examination of the liver is divided into the arterial phase (AP), the portal venous phase (VP) and the delayed phase (DP).
[0065] However, when the contrast agent is in different moments, the density and morphology of different tissues and organs in the scanned images are different. The portal vein in the liver is high-density in the VP phase and is basically not visible in the AP phase. The hepatic artery in the liver is high-density in the AP phase and is basically not visible in the VP phase. Due to the asymmetry of the intensity information of different types of blood vessels at different times, after the VP phase image is registered with the AP phase image, the portal vein and hepatic artery may have abnormalities in anatomical position and structure, resulting in vascular interpenetration and vascular malformations.
[0066] In the related art, in order to reduce the risk of vascular interpenetration, the method mainly finds out the possible interpenetrating blood vessels from the blood vessels in different phase images based on the position information of the blood vessels, and processes the possible interpenetrating blood vessels by inverting the colors, so as to reduce the vascular interpenetration phenomenon in the image registration process. However, in the process of image registration, this method needs to use strong constraint parameters to separate the interpenetrating blood vessels, which will cause obvious distortion of the vascular morphology. The above constraint parameters refer to some parameters that affect the registration process. Different combinations of parameters will affect the registration deformation field, and the influence is presented on the vascular image as vascular interpenetration or morphological distortion. Among them, the constraint parameters may include at least one of the registration model adopted by the registration algorithm, the optimization method adopted by the registration algorithm, the regularization strength of the registration parameters, the smoothing coefficient of the registration parameters, the iterative step size of the registration parameters, and the number of iterations of the registration parameters. For example, when the constraint parameter is 1, the initial value of a pixel point in the blood vessel is 1, and the result after the initial value is registered is 2. When the true value is between 1 and 2, the vascular morphology will be significantly distorted.
[0067] In order to ensure that blood vessels do not intersect during the multi-phase image registration process of the liver and maintain normal vascular morphology, this application provides an image registration method, which solves the problem of vascular morphological distortion while reducing vascular intersect overlap by locally rigidly adjusting the elastic displacement field of blood vessels that may intersect. Next, the image registration method provided by this application is introduced in detail.
[0068] The image registration method provided in the embodiment of the present application can be applied to Figure 1In the application environment shown. For example, the computer device may be a server, a personal computer, a laptop, a smart phone, a tablet computer, a smart mobile phone, etc. The computer device may include a processor, a memory and a network interface connected by a system bus or wirelessly. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device may include a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store data in the image registration process. The network interface of the computer device is used to communicate with an external terminal through a network connection, and the computer program is executed by the processor to implement an image registration method. Among them, the computer device can be implemented by an independent computer device or a computer device cluster composed of multiple computer devices. It should be noted that the memory of the computer device is not limited to the above-mentioned memory, and may also include a high-speed random access memory, a volatile solid-state memory, etc. In addition, the composition architecture of the computer device is not limited to the above-mentioned situation, and some components may be added or omitted.
[0069] In one embodiment, Figure 2 As shown, an image registration method is provided, which is applied to Figure 1 The computer device in the example is used to illustrate, including the following steps:
[0070] S201, obtaining an initial registration displacement field between a reference image and a floating image of a target part.
[0071] The target part can be any part of the simulated human body. Taking the liver as an example, the reference image and floating image of the liver can be examination images of the liver at different stages, the reference image can be a portal venous phase image, and the floating image can be an arterial phase image. The above registration displacement field refers to the registration direction and displacement between the pixel points of the reference image and the pixel points of the floating image.
[0072] In this embodiment, the computer device can search for two images of the target part at different times from the image database, and use any one of the two images as the reference image of the target part, and the other as the floating image of the target part. Then, the reference image and the floating image of the target part are registered using an image registration algorithm to obtain an initial registration displacement field between the reference image and the floating image. The image registration algorithm can be a correlation coefficient method, a dynamic programming method, a projection change method, etc.
[0073] S202, based on the initial registration displacement field, obtaining an elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part.
[0074] In this embodiment, after obtaining the initial registration displacement field between the reference image and the floating image, the computer device can perform rough segmentation on the reference image and the floating image respectively to obtain the reference image tubular structure mask corresponding to the reference image and the floating image tubular structure mask corresponding to the floating image. Then, the computer device can determine the registration displacement field corresponding to the tubular structure mask from the initial registration displacement field according to the reference image tubular structure mask and the floating image tubular structure mask, and determine the registration displacement field as the elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask.
[0075] S203 , based on the overlapping area between the tubular structure mask of the reference image and the tubular structure mask of the floating image, rigidly adjust the elastic registration displacement field to obtain a rigid displacement field of the overlapping area.
[0076] The rigid adjustment refers to dividing the elastic registration displacement field into a plurality of local registration displacement fields, and performing a uniform rigid adjustment on each local registration displacement field.
[0077] In this embodiment, the computer device can fuse the reference image tubular structure mask with the floating image tubular structure mask, and obtain the intersection area of the tubular structure mask in the fused image, and determine the intersection area as the overlapping area. Based on the overlapping area, the elastic registration displacement field corresponding to the overlapping area can be determined from the elastic registration displacement field, and the elastic registration displacement field corresponding to the overlapping area can be divided into a plurality of local elastic registration displacement fields. Afterwards, the computer device can convert each local elastic registration displacement field into a corresponding local rigid displacement field using a model transformation estimation method or a statistical method, and determine the converted local rigid displacement field as the rigid displacement field of the overlapping area. Among them, the model transformation method can include a random sampling consensus (RANSAC) algorithm, a principal component analysis method (PCA) or a linear regression or a deep learning method, etc. The statistical method can include a mean value, a median, a vector sum or a vector difference, etc.
[0078] S204, obtaining an image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field.
[0079] In this embodiment, after the rigid displacement field of the overlapping area is obtained, the computer device can determine the initial registration displacement field of the overlapping area from the initial registration displacement field based on the overlapping area. And the initial registration displacement field of the overlapping area is replaced by the rigid displacement field of the overlapping area, so that the target registration displacement field between the reference image and the floating image can be obtained. Therefore, the image registration result of the reference image and the floating image can be obtained based on the target registration displacement field of the reference image and the floating image. The displacement field of the overlapping area in the image registration process is a displacement field after rigid transformation, and the tubular results with overlapping conditions are separated, and the obtained image registration result avoids the deformity of the tubular structure.
[0080] In the above image registration method, an initial registration displacement field between a reference image and a floating image of a target part is obtained; based on the initial registration displacement field, an elastic registration displacement field between a tubular structure mask of the reference image and a tubular structure mask of the floating image of the target part is obtained; based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image, the elastic registration displacement field is rigidly adjusted to obtain a rigid displacement field of the overlapping area; based on the rigid displacement field of the overlapping area and the initial registration displacement field, an image registration result of the reference image and the floating image is obtained. The method uses the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image to rigidly adjust the elastic registration displacement field between the tubular structure mask of the reference image and the tubular structure mask of the floating image, so that the overlapping tubular structures can be separated and processed, thereby avoiding the morphological distortion of the tubular structures in the registration process due to the overlap. Based on the adjusted rigid displacement field of the overlapping area and the initial registration displacement field, the problem of morphological distortion of the tubular structure in the image registration result of the reference image and the floating image is avoided, thereby improving the accuracy of image registration.
[0081] Based on the above embodiments, this embodiment is to Figure 2 The relevant contents of step S203 "rigidly adjusting the elastic registration displacement field based on the overlapping area of the tubular structure mask of the reference image and the tubular structure mask of the floating image to obtain the rigid displacement field of the overlapping area" are introduced and explained. Figure 3 As shown, as a non-limiting example, the above step S203 may include the following content:
[0082] S301, determining the elastic registration displacement field to be adjusted corresponding to the overlapping area from the elastic registration displacement field.
[0083] In this embodiment, after the overlapping area is obtained, the computer device can determine the elastic registration displacement field of the overlapping area from the elastic registration displacement field according to the mapping relationship between the registration displacement field and the image tubular structure mask, and determine the elastic registration displacement field of the overlapping area as the elastic registration displacement field to be adjusted. It should be noted that the number of overlapping areas is at least one.
[0084] S302 , performing transformation estimation on the elastic registration displacement field to be adjusted to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted.
[0085] In this embodiment, the computer device can divide the elastic registration displacement field to be adjusted according to a preset division rule to obtain multiple local elastic registration displacement fields to be adjusted. For any local elastic registration displacement field, the computer device can fit each local elastic registration displacement field to obtain an estimated transformation model corresponding to the local elastic registration displacement field.
[0086] S303, using the estimated transformation model, rigidly adjusting the elastic registration displacement field to be adjusted to obtain a rigid displacement field of the overlapping area.
[0087] In this embodiment, for any local elastic registration displacement field to be adjusted, the computer device can use the estimated transformation model corresponding to the local elastic registration displacement field to be adjusted to perform rigid adjustment on the local elastic registration displacement field to be adjusted, so that all points of the local elastic registration displacement field to be adjusted satisfy the estimated transformation model. The above rigid adjustment operation is performed on all the local elastic registration displacement fields to be adjusted to obtain the rigid displacement field of the overlapping area.
[0088] In the above-mentioned image registration method, the elastic registration displacement field to be adjusted corresponding to the overlapping area is determined from the elastic registration displacement field; the elastic registration displacement field to be adjusted is transformed and estimated to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted; and the elastic registration displacement field to be adjusted is rigidly adjusted using the estimated transformation model to obtain a rigid displacement field of the overlapping area. This method can accurately obtain the estimated transformation model corresponding to the elastic registration displacement field to be adjusted by transforming and estimating the elastic registration displacement field to be adjusted corresponding to the overlapping area. Therefore, based on the estimated transformation model, the elastic registration displacement field to be adjusted can be rigidly adjusted more accurately, so that the obtained rigid displacement field of the overlapping area is more accurate.
[0089] Based on the above embodiments, this embodiment is to Figure 3 The relevant contents of step S302 "transform estimation of the elastic registration displacement field to be adjusted to obtain the estimated transformation model corresponding to the elastic registration displacement field to be adjusted" are introduced and explained. Figure 4As shown, as a non-limiting example, the above step S302 may include the following content:
[0090] S401 , dividing the elastic registration displacement field to be adjusted to obtain a plurality of local elastic registration displacement fields to be adjusted.
[0091] In this embodiment, the computer device can divide the elastic registration displacement field to be adjusted according to a preset division rule to obtain multiple local elastic registration displacement fields to be adjusted. The preset division rule can be obtained based on historical experience or can be divided based on a certain displacement field threshold. The sizes of the multiple local elastic registration displacement fields to be adjusted can be the same or different.
[0092] S402, selecting a plurality of initial displacement vectors from each local elastic registration displacement field to be adjusted to obtain an initial estimated transformation model.
[0093] In this embodiment, for any local elastic registration displacement field to be adjusted, the computer device can arbitrarily select multiple initial displacement vectors from the local elastic registration displacement field to be adjusted. Linear fitting is performed on each initial displacement vector to obtain a model corresponding to the initial displacement vector, and the model is determined as the initial estimated transformation model corresponding to the local elastic registration displacement field to be adjusted. Based on the above method, the initial estimated transformation model corresponding to each local elastic registration displacement field to be adjusted can be obtained.
[0094] S403, performing multiple iterations on each initial estimated transformation model until each initial estimated transformation model meets a preset condition, thereby obtaining an estimated transformation model corresponding to the elastic registration displacement field to be adjusted.
[0095] In this embodiment, for any initial estimated transformation model, since the initial estimated transformation model can only represent the transformation relationship of multiple initial displacement vectors, it cannot represent the transformation relationship of other displacement vectors in the local elastic registration displacement field to be adjusted. Therefore, the computer device can iterate the initial estimated transformation model multiple times so that the iterated initial estimated transformation model can represent the transformation relationship of all displacement vectors in the local elastic registration displacement field to be adjusted. When all the initial estimated models complete the iteration process, all the initial estimated models are determined as the estimated transformation models corresponding to the elastic registration displacement field to be adjusted.
[0096] In the above-mentioned image registration method, the elastic registration displacement field to be adjusted is divided to obtain a plurality of local elastic registration displacement fields to be adjusted; a plurality of initial displacement vectors are selected from each local elastic registration displacement field to be adjusted to obtain an initial estimated transformation model; each initial estimated transformation model is iterated multiple times until each initial estimated transformation model meets a preset condition, thereby obtaining an estimated transformation model corresponding to the elastic registration displacement field to be adjusted. The method can more accurately obtain the estimated transformation model corresponding to the elastic registration displacement field to be adjusted by selecting an initial displacement vector from each local elastic registration displacement field to be adjusted, determining the initial estimated transformation model corresponding to each local elastic registration displacement field to be adjusted, and iterating multiple times for each initial estimated transformation model.
[0097] Based on the above embodiments, this embodiment is to Figure 4 The relevant contents of step S402 "performing multiple iterations on each initial estimation transformation model" are introduced and explained. Figure 5 As shown, as a non-limiting example, the above step S402 may include the following content:
[0098] S501 , for any iteration process of any initial estimated transformation model, obtaining the projection errors between all displacement vectors in the local elastic registration displacement field to be adjusted and the estimated transformation model obtained in the previous iteration.
[0099] In this embodiment, during the process of iterating the initial estimated transformation model, the computer device can obtain the estimated transformation model obtained in the previous iteration, and project all displacement vectors in the local elastic registration displacement field to be adjusted onto the estimated transformation model, obtain the projection distance between each displacement vector and the estimated transformation model, and determine the projection distance as the projection error.
[0100] S502: taking a displacement vector whose projection error is less than a preset error threshold as an iterative displacement vector.
[0101] In this embodiment, the computer device may compare the projection error between each displacement vector and the estimated transformation model with a preset error threshold. If the projection error is less than the preset error threshold, it means that the projection distance between the displacement vector corresponding to the projection error and the estimated transformation model is small, and the partial displacement vector may be determined as an iterative displacement vector, and the estimated transformation model obtained by the previous iteration may be iterated using the iterative displacement vector. If the projection error is greater than or equal to the preset error threshold, it means that the projection distance between the displacement vector corresponding to the projection error and the estimated transformation model is large, and there is no need to pay attention to the partial displacement vector.
[0102] S503, iterating the estimated transformation model obtained in the previous iteration based on the iterative displacement vector.
[0103] In this embodiment, after obtaining the iterative displacement vector, the computer device can adjust the estimated transformation model obtained in the previous iteration based on the iterative displacement vector, the historical iterative displacement vector and the initial displacement vector, so that the adjusted estimated transformation model can represent the iterative displacement vector and the initial displacement vector. It should be noted that the historical iterative displacement vector represents the iterative displacement vector of each historical iteration in the historical iteration process. For example, if the current iteration is the 5th iteration, the 5th iterative displacement vector, the iterative displacement vectors of the previous 4 times and the initial displacement vector are used in the process of adjusting the estimated transformation model of the 4th iteration. In other words, the displacement vector used in each iteration process will increase.
[0104] In the above-mentioned image registration method, for any iteration process of any initial estimated transformation model, the projection error between all displacement vectors in the elastic registration displacement field to be adjusted and the estimated transformation model obtained in the previous iteration is obtained; the displacement vector whose projection error is less than the preset error threshold is used as the iterative displacement vector; based on the iterative displacement vector, the estimated transformation model obtained in the previous iteration is iterated. In the process of iterating the initial estimated transformation model of this method, each iteration will increase the displacement vector on the basis of the previous one. In this way, the initial estimated transformation model is more accurate in each iteration process than the previous iteration, and the obtained estimated transformation model will also be more accurate.
[0105] Based on the above embodiments, this embodiment is to Figure 3 The relevant contents of step S303 "using the estimated transformation model to rigidly adjust the elastic registration displacement field to be adjusted to obtain the rigid displacement field of the overlapping area" are introduced and explained. Figure 6 As shown, as a non-limiting example, the above step S303 may include the following content:
[0106] S601, using the estimated transformation model, performing local rigid transformation on the elastic registration displacement field to be adjusted to obtain multiple local rigid displacement fields.
[0107] In this embodiment, each local elastic registration displacement field to be adjusted corresponds to an estimated transformation model, and the computer device can determine the estimated transformation model corresponding to each local elastic registration displacement field to be adjusted based on the corresponding relationship. And use the estimated transformation model to perform local rigid transformation on the displacement vectors in the local elastic registration displacement field to be adjusted, so that all displacement vectors satisfy the estimated transformation model, and obtain the local rigid displacement field corresponding to the estimated transformation model. All the elastic registration displacement fields to be adjusted are adjusted in the above manner to obtain multiple local rigid displacement fields.
[0108] S602, fusing the local rigid displacement fields to obtain the rigid displacement field of the overlapping area.
[0109] In this embodiment, after acquiring multiple local rigid displacement fields, the computer device can use a fusion algorithm to fuse the multiple local rigid displacement fields, and determine the fused rigid displacement field as the rigid displacement field of the overlapping area. The fusion algorithm can be a Gaussian weighted method, which can eliminate the influence of edge effects. The fusion algorithm can also be a linear weighted fusion method, a feature fusion method, a cross fusion method, etc.
[0110] In the above-mentioned image registration method, the estimated transformation model is used to perform local rigid transformation on the elastic registration displacement field to be adjusted to obtain multiple local rigid displacement fields; the local rigid displacement fields are fused to obtain the rigid displacement field of the overlapping area. This method can perform rigid transformation on the elastic registration displacement field to be adjusted in a more targeted manner by means of local rigid transformation to obtain multiple more accurate local rigid displacement fields. Afterwards, the local rigid displacement fields are fused to obtain a higher accuracy rigid displacement field of the overlapping area.
[0111] Based on the above embodiments, this embodiment is to Figure 2 The relevant contents of step S202 "obtaining the elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part based on the initial registration displacement field" are introduced and explained. Figure 7 As shown, as a non-limiting example, the above step S202 may include the following content:
[0112] S701, performing sampling transformation on the tubular structure mask of the floating image using the initial registration displacement field.
[0113] The sampling transformation refers to converting multiple sampling points of the tubular structure mask of the floating image into the space of the reference image.
[0114] In this embodiment, the computer device may determine a plurality of sampling points in the tubular structure mask of the floating image, and transform the plurality of sampling points in the tubular structure mask of the floating image into the space of the reference image using the initial registration displacement field.
[0115] S702, performing color inversion processing on the tubular structure mask of the floating image and the tubular structure mask of the reference image after sampling transformation.
[0116] The inversion process refers to processing the colors of the tubular structure mask of the floating image after sampling transformation and the tubular structure mask of the reference image into colors with a large color difference. For example, the color of the tubular structure mask of the floating image after sampling transformation is processed to red, and the color of the tubular structure mask of the reference image is processed to green. Alternatively, the color value of the tubular structure mask of the floating image after sampling transformation can be set to 1, and the color value of the tubular structure mask of the reference image can be set to -1.
[0117] In this embodiment, in order to strongly distinguish the tubular structure mask of the floating image after sampling transformation from the tubular structure mask of the reference image, the computer device can set the colors of the tubular structure mask of the floating image after sampling transformation and the tubular structure mask of the reference image to opposite colors. In this way, the tubular structure mask of the floating image after sampling transformation and the tubular structure mask of the reference image can be distinguished more clearly and intuitively.
[0118] S703, elastically registering the tubular structure mask of the floating image after the inversion processing and the tubular structure mask of the reference image to obtain an elastically registered displacement field.
[0119] The elastic registration refers to the process of registering each pixel point in the tubular structure mask of the floating image with each pixel point in the tubular structure mask of the reference image.
[0120] In this embodiment, the computer device can use an image registration algorithm to elastically register the tubular structure mask of the floating image after inversion processing and the tubular structure mask of the reference image, obtain a displacement field between the tubular structure mask of the floating image after inversion processing and the tubular structure mask of the reference image, and determine the displacement field as the elastic registration displacement field.
[0121] In the above image registration method, the tubular structure mask of the floating image is sampled and transformed using the initial registration displacement field; the tubular structure mask of the floating image and the tubular structure mask of the reference image after the sampling and transformation are subjected to color inversion processing; the tubular structure mask of the floating image and the tubular structure mask of the reference image after the color inversion processing are elastically registered to obtain the elastic registration displacement field. Before the elastic registration is performed, the tubular structure mask of the floating image is converted to the space of the reference image by sampling and transformation, and the tubular structure mask of the floating image and the tubular structure mask of the reference image after the sampling and transformation are subjected to color inversion processing, so that the tubular structure mask of the floating image and the tubular structure mask of the reference image after the registration and the color inversion processing can be more clearly distinguished.
[0122] Based on the above embodiments, this embodiment is to Figure 2 The relevant contents of step S204 "obtaining the image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field" are introduced and explained. Figure 8As shown, as a non-limiting example, the above step S204 may include the following content:
[0123] S801, composite the rigid displacement field of the overlapping area with the initial registration displacement field to obtain a composite displacement field between the reference image and the floating image.
[0124] In this embodiment, after obtaining the rigid displacement field of the overlapping area, the computer device can determine the initial registration displacement field corresponding to the overlapping area based on the overlapping area, and use the rigid displacement field of the overlapping area to replace the initial registration displacement field corresponding to the overlapping area, so as to obtain a composite displacement field between the reference image and the floating image. The composite processing process can be expressed as:
[0125]
[0126] in, is the initial registration displacement field, is the rigid displacement field of the overlapping area, and x represents the coordinate of the reference image.
[0127] S802, obtaining an image registration result between the reference image and the floating image based on the composite displacement field.
[0128] In this embodiment, the computer device may perform image registration on the reference image and the floating image according to the composite displacement field between the reference image and the floating image to obtain an image registration result of the reference image and the floating image.
[0129] In the above image registration method, the rigid displacement field of the overlapping area is compounded with the initial registration displacement field to obtain a composite displacement field between the reference image and the floating image; based on the composite displacement field, the image registration result of the reference image and the floating image is obtained. This method can accurately obtain the image registration result of the reference image and the floating image by compounding the rigid displacement field of the overlapping area with the initial registration displacement field and based on the composite displacement field; at the same time, there is no deformity in the tubular structure in the image registration result.
[0130] Based on the above embodiment, this embodiment introduces and explains the contents related to obtaining the image registration result between the reference image and the floating image. Fig. 9 As shown, as a non-limiting example, the above method may also include the following content:
[0131] S901, obtaining a mask heat map corresponding to the tubular structure mask of the floating image; different probability ranges in the mask heat map correspond to different identification information.
[0132] The mask heat map is an intermediate segmentation result of the tubular structure obtained by the tubular structure segmentation algorithm. Taking the tubular structure as a blood vessel as an example, the mask heat map of the blood vessel is a continuous probability distribution map, which can provide more blood vessel information, such as the strength, morphology and location information of the blood vessel.
[0133] In this embodiment, the computer device can use a feature segmentation algorithm to extract and segment the floating image, determine the probability that each pixel in the floating image is a tubular structure mask, and divide the probability corresponding to each pixel into different probability ranges according to the probability level. Different probability ranges are represented by different identification information to obtain a mask heat map corresponding to the tubular structure mask of the floating image. For example, pixels with a probability value of 80%-100% are represented in red, and pixels with a probability value of 60%-80% are represented in blue. Among them, the feature segmentation algorithm can be a probability-based feature segmentation algorithm, a classifier-based segmentation algorithm, or a deep learning-based segmentation algorithm.
[0134] S902: performing interpolation sampling on the mask heat map based on the composite displacement field.
[0135] In this embodiment, the computer device can use a related interpolation method to perform interpolation sampling processing on the mask thermal map based on the composite displacement field to obtain a mask thermal map after difference sampling. For example, the related interpolation method can be Newton interpolation, piecewise linear interpolation, or cubic spline interpolation.
[0136] S903, obtaining an image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image according to the mask heat map after interpolation sampling and a preset probability threshold.
[0137] In this embodiment, after obtaining the mask heat map after interpolation sampling, the computer device can compare the probability of each sampling point on the mask heat map after interpolation sampling with the preset probability threshold. If the probability of a certain sampling point is greater than the preset threshold, the pixel point corresponding to the sampling point is determined as the image registration result of the tubular structure mask. It should be noted that if the mask heat map is not interpolated and sampled, and the mask heat map is directly divided using the preset probability threshold, it will lead to information loss and discontinuity. The division of the mask heat map can only determine whether the blood vessel exists, but lacks the strength and position information of the blood vessel edge. Therefore, the discontinuity of the tubular structure mask segmentation can be solved by the interpolation method.
[0138] In the above-mentioned image registration method, a mask heat map corresponding to the tubular structure mask of the floating image is obtained; the mask heat map is interpolated and sampled based on the composite displacement field; and the image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image is obtained according to the interpolated and sampled mask heat map and a preset probability threshold. This method performs interpolation sampling processing on the mask heat map based on the composite displacement field, and the interpolation sampling can more accurately estimate the strength and position information of the edge of the blood vessel, so that the obtained blood vessel morphology is smoother and the interpolation result is more accurate.
[0139] As a specific embodiment of the present application, Fig.10 As shown, the image registration method includes:
[0140] S1001, obtaining an initial registration displacement field between a reference image and a floating image of a target part;
[0141] S1002, performing sampling transformation on the tubular structure mask of the floating image using the initial registration displacement field;
[0142] S1003, performing color inversion processing on the tubular structure mask of the floating image and the tubular structure mask of the reference image after sampling transformation;
[0143] S1004, elastically registering the tubular structure mask of the floating image after the inversion processing and the tubular structure mask of the reference image to obtain an elastically registered displacement field;
[0144] S1005, determining the elastic registration displacement field to be adjusted corresponding to the overlapping area from the elastic registration displacement field;
[0145] S1006, dividing the elastic registration displacement field to be adjusted to obtain a plurality of local elastic registration displacement fields to be adjusted;
[0146] S1007, selecting a plurality of initial displacement vectors from each local elastic registration displacement field to be adjusted to obtain an initial estimated transformation model;
[0147] S1008, for any iteration process of any initial estimated transformation model, obtaining the projection errors between all displacement vectors in the local elastic registration displacement field to be adjusted and the estimated transformation model obtained in the previous iteration;
[0148] S1009, taking a displacement vector whose projection error is less than a preset error threshold as an iterative displacement vector;
[0149] S1010, iterating the estimated transformation model obtained in the previous iteration based on the iterative displacement vector until the initial estimated transformation model meets a preset condition, thereby obtaining an estimated transformation model corresponding to the elastic registration displacement field to be adjusted;
[0150] S1011, using the estimated transformation model, performing a local rigid transformation on the elastic registration displacement field to be adjusted to obtain a plurality of local rigid displacement fields;
[0151] S1012, fusing each local rigid displacement field to obtain a rigid displacement field in the overlapping area;
[0152] S1013, combining the rigid displacement field of the overlapping area with the initial registration displacement field to obtain a composite displacement field between the reference image and the floating image;
[0153] S1014, obtaining an image registration result between the reference image and the floating image based on the composite displacement field;
[0154] S1015, obtaining a mask heat map corresponding to the tubular structure mask of the floating image;
[0155] S1016, performing interpolation sampling on the mask thermal map based on the composite displacement field;
[0156] S1017, obtaining an image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image according to the mask heat map after interpolation sampling and a preset probability threshold.
[0157] Fig.11 The figure is a flow chart of the image registration method. The tubular structure in the target part is a blood vessel. First, the reference image and the floating image are initially registered to obtain an initial registration displacement field. Based on the initial registration displacement field, the elastic registration displacement field between the floating image vascular mask and the reference image vascular mask is obtained. At the same time, the reference image vascular mask and the floating image vascular mask are fused to obtain the vascular mask of the overlapping area. The elastic registration displacement field of the overlapping area is determined from the elastic registration field, and the elastic registration displacement of the overlapping area is rigidly transformed to obtain the rigid displacement field of the overlapping area. Then, the initial registration displacement field and the rigid displacement field of the overlapping area are composited to obtain a composite displacement field. Then, the floating image vascular mask is interpolated and sampled to obtain the image registration result of the floating image vascular mask and the reference image vascular mask.
[0158] Figure 12-Figure 15 This is a schematic diagram of the vascular image registration results. The darker part in the figure is the overlapping area of the two images. Fig.12 , 14 To directly perform image registration, Fig.13 , Fig.15 The result of image registration by the technical solution in this application is Fig.12 and Fig.13 For comparison, and Fig.14 and Fig.15In contrast, the technical solution in the present application can reduce the problem of morphological distortion of blood vessels in the registration results, thereby improving the accuracy of image registration.
[0159] It should be understood that, although the steps in the flowcharts involved in the above embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.
[0160] Based on the same inventive concept, the embodiment of the present application also provides an image registration device for implementing the above-mentioned image registration method. The implementation solution provided by the device to solve the problem is similar to the implementation solution recorded in the above-mentioned method, so the specific limitations in one or more image registration device embodiments provided below can refer to the limitations of the image registration method above, and will not be repeated here.
[0161] In one embodiment, Fig.16 As shown, an image registration device is provided, comprising: a first acquisition module 11, a second acquisition module 12, an adjustment module 13 and a determination module 14, wherein:
[0162] A first acquisition module 11 is used to acquire an initial registration displacement field between a reference image and a floating image of a target part;
[0163] A second acquisition module 12 is used to acquire an elastic registration displacement field between a reference image tubular structure mask and a floating image tubular structure mask of a target part based on the initial registration displacement field;
[0164] An adjustment module 13, configured to perform a rigid adjustment on the elastic registration displacement field based on an overlapped area between the tubular structure mask of the reference image and the tubular structure mask of the floating image, so as to obtain a rigid displacement field of the overlapped area;
[0165] The determination module 14 is used to obtain the image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field.
[0166] In one embodiment, the adjustment module 13 includes: a first determination unit, a transformation estimation unit and an adjustment unit, wherein:
[0167] A first determining unit, configured to determine the elastic registration displacement field to be adjusted corresponding to the overlapping area from the elastic registration displacement field;
[0168] A transformation estimation unit, used for performing transformation estimation on the elastic registration displacement field to be adjusted, and obtaining an estimated transformation model corresponding to the elastic registration displacement field to be adjusted;
[0169] The adjustment unit is used to use the estimated transformation model to perform rigid adjustment on the elastic registration displacement field to be adjusted, so as to obtain the rigid displacement field of the overlapping area.
[0170] In one embodiment, the transformation estimation unit is also used to divide the elastic registration displacement field to be adjusted to obtain multiple local elastic registration displacement fields to be adjusted; select multiple initial displacement vectors from each local elastic registration displacement field to be adjusted to obtain an initial estimated transformation model; and iterate each initial estimated transformation model multiple times until each initial estimated transformation model meets a preset condition to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted.
[0171] In one embodiment, the transformation estimation unit is also used to obtain the projection error between all displacement vectors in the local elastic registration displacement field to be adjusted and the estimated transformation model obtained in the previous iteration for any iteration process of any initial estimated transformation model; take the displacement vector whose projection error is less than a preset error threshold as the iterative displacement vector; and iterate the estimated transformation model obtained in the previous iteration based on the iterative displacement vector.
[0172] In one embodiment, the adjustment unit is further used to use the estimated transformation model to perform local rigid transformation on the elastic registration displacement field to be adjusted to obtain multiple local rigid displacement fields; and fuse the local rigid displacement fields to obtain the rigid displacement field of the overlapping area.
[0173] In one embodiment, the second acquisition module includes: a transformation unit, a color inversion processing unit and a registration unit, wherein:
[0174] A transformation unit, used for performing sampling transformation on the tubular structure mask of the floating image by using the initial registration displacement field;
[0175] A color inversion processing unit, used for performing color inversion processing on the tubular structure mask of the floating image and the tubular structure mask of the reference image after sampling transformation;
[0176] The registration unit is used to elastically register the tubular structure mask of the floating image after the inversion processing and the tubular structure mask of the reference image to obtain an elastic registration displacement field.
[0177] In one embodiment, the determination module includes: a compound unit and a second determination unit, wherein:
[0178] A compounding unit, used for compounding the rigid displacement field of the overlapping area with the initial registration displacement field to obtain a compound displacement field between the reference image and the floating image;
[0179] The second determining unit is used to obtain an image registration result between the reference image and the floating image based on the composite displacement field.
[0180] In one embodiment, the above-mentioned image registration device further includes an acquisition unit, a sampling unit and a third determination unit, wherein:
[0181] An acquisition unit is used to acquire a mask heat map corresponding to the tubular structure mask of the floating image; different probability ranges in the mask heat map correspond to different identification information;
[0182] A sampling unit, used for interpolating and sampling the mask thermal map based on the composite displacement field;
[0183] The third determination unit is used to obtain an image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image according to the mask heat map after interpolation sampling and a preset probability threshold.
[0184] Each module in the above-mentioned image registration device can be implemented in whole or in part by software, hardware or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each module above.
[0185] 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, any one of the contents in the above-mentioned image registration method embodiments is implemented.
[0186] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, any one of the contents in the above-mentioned image registration method embodiments is implemented.
[0187] In one embodiment, a computer program product is provided, including a computer program, which implements any one of the contents of the above-mentioned image registration method embodiments when executed by a processor.
[0188] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.
[0189] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed 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 the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. 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). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but are not limited to this.
[0190] The technical features of the above embodiments may be arbitrarily combined. 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.
[0191] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.
Claims
1. An image registration method, characterized in that: The method comprises: obtaining an initial registration displacement field between the reference image and the floating image of the target part; Based on the initial registration displacement field, obtaining an elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part; Based on the overlapping area between the tubular structure mask of the reference image and the tubular structure mask of the floating image, rigidly adjusting the elastic registration displacement field to obtain a rigid displacement field of the overlapping area; An image registration result between the reference image and the floating image is obtained based on the rigid displacement field of the overlapping area and the initial registration displacement field.
2. The method according to claim 1, characterized in that The step of rigidly adjusting the elastic registration displacement field based on the overlapping area between the reference image tubular structure mask and the floating image tubular structure mask to obtain the rigid displacement field of the overlapping area includes: Determining the elastic registration displacement field to be adjusted corresponding to the overlapping area from the elastic registration displacement field; Performing transformation estimation on the elastic registration displacement field to be adjusted to obtain an estimated transformation model corresponding to the elastic registration displacement field to be adjusted; The estimated transformation model is used to perform rigid adjustment on the elastic registration displacement field to be adjusted to obtain the rigid displacement field of the overlapping area.
3. The method according to claim 2, characterized in that The step of using the estimated transformation model to rigidly adjust the elastic registration displacement field to be adjusted to obtain the rigid displacement field of the overlapping area includes: Using the estimated transformation model, performing a local rigid transformation on the elastic registration displacement field to be adjusted to obtain a plurality of local rigid displacement fields; The local rigid displacement fields are fused to obtain the rigid displacement field of the overlapping area.
4. The method according to any one of claims 1 to 3, characterized in that: The step of obtaining the elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part based on the initial registration displacement field comprises: Performing sampling transformation on the tubular structure mask of the floating image using the initial registration displacement field; Performing color inversion processing on the tubular structure mask of the floating image after sampling transformation and the tubular structure mask of the reference image; The tubular structure mask of the floating image after the inversion process and the tubular structure mask of the reference image are elastically registered to obtain the elastic registration displacement field.
5. The method according to any one of claims 1 to 3, characterized in that: The obtaining of the image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field comprises: Combining the rigid displacement field of the overlapping area with the initial registration displacement field to obtain a composite displacement field between the reference image and the floating image; Based on the composite displacement field, an image registration result between the reference image and the floating image is obtained.
6. The method according to claim 5, characterized in that The method further comprises: Acquire a mask heat map corresponding to the tubular structure mask of the floating image; different probability ranges in the mask heat map correspond to different identification information; Performing interpolation sampling on the mask thermal map based on the composite displacement field; According to the mask heat map after interpolation sampling and a preset probability threshold, an image registration result between the tubular structure mask of the floating image and the tubular structure mask of the reference image is obtained.
7. An image registration device, characterized in that: The device comprises: A first acquisition module is used to acquire an initial registration displacement field between a reference image and a floating image of a target part; A second acquisition module is used to acquire an elastic registration displacement field between the reference image tubular structure mask and the floating image tubular structure mask of the target part based on the initial registration displacement field; An adjustment module, configured to perform a rigid adjustment on the elastic registration displacement field based on an overlapped area between the reference image tubular structure mask and the floating image tubular structure mask, so as to obtain a rigid displacement field of the overlapped area; The determination module is used to obtain the image registration result between the reference image and the floating image based on the rigid displacement field of the overlapping area and the initial registration displacement field.
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 6 are implemented.
9. A computer-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 6 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 6 are implemented.