2D / 3D medical image registration method, system, device and medium

Abstract

The invention discloses a 2D / 3D medical image registration method, system and device and a medium. The method comprises the steps that a normal X-ray film and a normal DRR are acquired, and a lateral X-ray film and a lateral DRR are acquired; calculating the similarity between the normal position X-ray film and the normal position DRR, and calculating the similarity between the lateral position X-ray film and the lateral position DRR; carrying out weighted summation on the similarity between the normal position X-ray film and the normal position DRR and the similarity between the lateral position X-ray film and the normal position DRR to obtain similarity measurement; and by taking the registration parameters as variables, optimizing the similarity measurement result through a particle swarm algorithm, so that the similarity measurement reaches an extreme value, and obtaining the registration parameters corresponding to the similarity measurement reaching the extreme value, the method, the system, the equipment and the medium can improve the precision of 2D / 3D medical image registration.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing and relates to a 2D / 3D medical image registration method, system, equipment and medium. Background Art

[0002] In the registration phase of minimally invasive spinal surgery, 2D / 3D medical image registration algorithms based on image grayscale are widely used. However, different X-ray imaging devices have different imaging styles and the X-ray films generated by different X-ray imaging devices may have problems such as blurred imaging and low contrast, which will lead to low registration accuracy of the registration algorithm and cannot meet the accuracy requirements of minimally invasive spinal surgery for the registration phase.

[0003] In surgical navigation treatment, 2D / 3D image registration is a key technology that can assist doctors in accurately locating lesions. It is widely used in image-guided minimally invasive surgery, radiotherapy planning, and postoperative efficacy evaluation and other interventional treatment fields. With the continuous development of minimally invasive surgery technology, minimally invasive spine surgery has become one of the important directions of spinal surgery. This type of surgery requires high-precision navigation technology to assist doctors in operation. Three-dimensional image data such as CT, due to its intuitiveness and presentation of three-dimensional information of patients, can help doctors accurately locate the lesion area and formulate a thorough surgical plan. Although three-dimensional image data such as CT can provide rich spatial information, it cannot be obtained in real time during surgery. In contrast, two-dimensional images such as X-ray images can be obtained in real time during surgery, but lack complete three-dimensional spatial information. Therefore, by combining intraoperative two-dimensional X-ray images with preoperative CT three-dimensional data for registration, the lack of spatial information can be effectively compensated and the advantages of both can be comprehensively utilized.

[0004] At present, many scholars have conducted extensive research on 2D-3D medical image registration algorithms, which are generally divided into two categories: feature-based methods and grayscale-based methods. Feature-based registration methods usually rely on the spatial positioning of external markers. Before surgery, markers or feature points are placed on the patient's skin surface or implanted in the body. During surgery, the spatial position of these markers is tracked by detection algorithms or detection equipment to achieve accurate positioning of the lesion. Feature-based methods have low computational complexity and simple implementation. The spatial information of the lesion can be obtained through markers, and the requirements are low. However, it relies on the precise placement of markers, soft tissue deformation will cause errors, and it is only applicable to rigid registration with small displacement changes. Grayscale-based image registration methods usually achieve registration by analyzing the grayscale relationship between pixels. Its core principle is to use a certain optimization algorithm to gradually reduce the difference in grayscale values ​​between two images in iterative calculations. With each iteration, the similarity of the images will gradually increase until the difference between the two images is reduced to a certain extent, thereby achieving the final registration effect. In this way, images can be matched more accurately, especially in the absence of obvious feature points, grayscale registration is particularly effective. Generally speaking, the grayscale-based image registration method mainly includes three parts: DRR image generation, similarity measurement, and iterative method. First, the ray projection method is used to generate the anteroposterior DRR images from the preoperative CT data, and then the similarity measurement between the generated DRR images and the anteroposterior X-ray films taken by the intraoperative X-ray equipment is calculated. The difference between the calculated similarity measurement result and the threshold is compared. If the threshold requirement is not met, the initial parameters are changed through the iterative method to regenerate the anteroposterior DRR. If the threshold requirement is met, the algorithm is stopped and the algorithm operation result at this time is output as the registration parameter.

[0005] Traditional 2D / 3D medical image registration algorithms use the correlation of image pixel grayscale as the basis for registration in the design of objective functions. Different X-ray imaging devices have different imaging styles and X-ray films generated by different X-ray imaging devices may have problems such as blurred imaging and low contrast, which leads to unstable pixel grayscale information of X-ray films and seriously affects the accuracy of the registration algorithm. Summary of the invention

[0006] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and provide a 2D / 3D medical image registration method, system, device and medium, which can improve the accuracy of 2D / 3D medical image registration.

[0007] In order to achieve the above object, the present invention adopts the following technical scheme:

[0008] In one aspect, the present invention provides a 2D / 3D medical image registration method, comprising:

[0009] Get an AP X-ray and AP DRR, get a lateral X-ray and lateral DRR;

[0010] The similarity between the anteroposterior X-ray and the anteroposterior DRR was calculated, and the similarity between the lateral X-ray and the lateral DRR was calculated;

[0011] The similarity between the anteroposterior X-ray and the anteroposterior DRR and the similarity between the lateral X-ray and the anteroposterior DRR are weighted summed to obtain the similarity measure;

[0012] Taking the registration parameter as a variable, the similarity measure result is optimized by a particle swarm algorithm so that the similarity measure reaches an extreme value, and the registration parameter corresponding to the extreme value of the similarity measure is obtained.

[0013] The 2D / 3D medical image registration method of the present invention is further improved in that:

[0014] Furthermore, the process of obtaining the frontal X-ray film and the frontal DRR, and obtaining the lateral X-ray film and the lateral DRR is:

[0015] Take anteroposterior and lateral X-rays of the patient during the operation;

[0016] The patient's spine CT data taken before surgery is obtained, and the lateral DRR and anteroposterior DRR of the CT are generated using the ray projection method of the ITK library based on the initial alignment parameters and the patient's spine CT data.

[0017] Furthermore, before calculating the similarity between the frontal X-ray film and the frontal DRR and calculating the similarity between the lateral X-ray film and the lateral DRR, the step further includes:

[0018] The opencv library of python is used to denoise the anteroposterior and lateral X-rays.

[0019] Furthermore, the process of calculating the similarity between the frontal X-ray film and the frontal DRR is:

[0020] Calculate the normalized mutual information between the AP X-ray and the AP DRR;

[0021] Calculate the gradient-related information between the AP X-ray and AP DRR;

[0022] The normalized mutual information between the orthotopic X-ray film and the orthotopic DRR and the gradient-related information between the orthotopic X-ray film and the orthotopic DRR are weightedly summed to obtain the similarity between the orthotopic X-ray film and the orthotopic DRR.

[0023] Furthermore, the process of calculating the similarity between the lateral X-ray film and the lateral DRR is:

[0024] Calculate the normalized mutual information between the lateral X-ray and the lateral DRR;

[0025] Calculate the gradient-related information between the lateral X-ray and the lateral DRR;

[0026] The normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR are weightedly summed to obtain the similarity between the lateral X-ray film and the lateral DRR.

[0027] In a second aspect of the present invention, the present invention provides a 2D / 3D medical image registration system, comprising:

[0028] An acquisition module, used for acquiring an anteroposterior X-ray film and an anteroposterior DRR, and acquiring a lateral X-ray film and a lateral DRR;

[0029] The first calculation module is used to calculate the similarity between the frontal X-ray film and the frontal DRR, and calculate the similarity between the lateral X-ray film and the lateral DRR;

[0030] The second calculation module is used to perform weighted summation on the similarity between the frontal X-ray film and the frontal DRR and the similarity between the lateral X-ray film and the frontal DRR to obtain a similarity measure;

[0031] The optimization module is used to optimize the similarity measurement result by particle swarm algorithm with the registration parameter as a variable, so that the similarity measurement reaches an extreme value, and obtains the registration parameter corresponding to the extreme value of the similarity measurement.

[0032] The 2D / 3D medical image registration system of the present invention is further improved in that:

[0033] Furthermore, the first calculation module includes:

[0034] The third calculation module is used to calculate the normalized mutual information between the anteroposterior X-ray film and the anteroposterior DRR;

[0035] A fourth calculation module is used to calculate the gradient-related information of the anteroposterior X-ray film and the anteroposterior DRR;

[0036] The first weighted summation module is used to perform weighted summation on the normalized mutual information between the orthostatic X-ray film and the orthostatic DRR and the gradient-related information between the orthostatic X-ray film and the orthostatic DRR to obtain the similarity between the orthostatic X-ray film and the orthostatic DRR.

[0037] Furthermore, the first calculation module includes:

[0038] a fifth calculation module, for calculating the normalized mutual information between the lateral X-ray film and the lateral DRR;

[0039] A sixth calculation module, used to calculate the gradient-related information of the lateral X-ray film and the lateral DRR;

[0040] The second weighted summation module is used to perform weighted summation on the normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR to obtain the similarity between the lateral X-ray film and the lateral DRR.

[0041] In a third aspect of the present invention, the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the 2D / 3D medical image registration method when executing the computer program.

[0042] In a fourth aspect of the present invention, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the 2D / 3D medical image registration method are implemented.

[0043] The present invention has the following beneficial effects:

[0044] During specific operation, the 2D / 3D medical image registration method, system, device and medium described in the present invention obtain an anteroposterior X-ray and an anteroposterior DRR, obtain a lateral X-ray and a lateral DRR, and use the anteroposterior X-ray, anteroposterior DRR, lateral X-ray and lateral DRR as the basis for registration, that is, the image pixel grayscale information and the target object contour gradient information are simultaneously used as the registration basis, thereby eliminating the influence of inconsistent X-ray style, blurred imaging and low contrast on the accuracy of the registration algorithm, improving the robustness of the registration algorithm, and meeting the accuracy requirements of minimally invasive spinal surgery for the registration link. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0046] Figure 1 is a flowchart of the similarity measure calculation process;

[0047] Figure 2 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only an embodiment of a part of the present invention, not all embodiments, and is not intended to limit the scope of the present invention. In addition, in the following description, the description of well-known structures and technologies is omitted to avoid unnecessary confusion of the concepts disclosed in the present invention. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present invention.

[0049] The accompanying drawings show schematic diagrams of structures according to embodiments disclosed in the present invention. These figures are not drawn to scale, and some details are magnified and some details may be omitted for the purpose of clear expression. The shapes of various regions and layers shown in the figures and the relative sizes and positional relationships therebetween are only exemplary, and may deviate in practice due to manufacturing tolerances or technical limitations, and those skilled in the art may additionally design regions / layers with different shapes, sizes, and relative positions according to actual needs.

[0050] Embodiment 1

[0051] refer to Figure 1 and Figure 2 The 2D / 3D medical image registration method of the present invention comprises the following steps:

[0052] 1) During the operation, the patient's anteroposterior X-ray and lateral X-ray are taken, uploaded to the computer, and then the anteroposterior X-ray and lateral X-ray are denoised using the python opencv library to achieve preprocessing. The operator inputs the registration parameters of the initial position according to the empirical rules.

[0053] 2) Obtain the patient's spinal CT data taken before surgery, and generate the lateral DRR and anteroposterior DRR of the CT using the ray projection method of the ITK library based on the initial registration parameters and the patient's spinal CT data;

[0054] 3) Calculate the similarity between the frontal X-ray and the frontal DRR, and calculate the similarity between the lateral X-ray and the lateral DRR;

[0055] Specifically, the objective function is normalized mutual information (NMI) combined with gradient correlation (GC). First, the normalized mutual information is calculated using the registration_method method of the SITK library, and then the gradient-related information is calculated using the SITK library and the numpy library. The results obtained by both are numbers in the interval [-1, 1].

[0056] The process of calculating the similarity between the positive X-ray film and the positive DRR is as follows:

[0057] 311) The normalized mutual information between the anteroposterior X-ray and the anteroposterior DRR was calculated using the registration_method method of the SITK library;

[0058] 312) Use SITK library and numpy library to calculate the gradient related information of the anteroposterior X-ray film and anteroposterior DRR;

[0059] 313) The normalized mutual information between the anteroposterior X-ray film and the anteroposterior DRR and the gradient-related information between the anteroposterior X-ray film and the anteroposterior DRR are weightedly summed to obtain the similarity between the anteroposterior X-ray film and the anteroposterior DRR.

[0060] The process of calculating the similarity between the lateral X-ray film and the lateral DRR is as follows:

[0061] 321) The normalized mutual information between lateral X-ray and lateral DRR was calculated by the registration_method method of the SITK library;

[0062] 322) Use SITK library and numpy library to calculate the gradient related information of lateral X-ray and lateral DRR;

[0063] 323) The normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR are weightedly summed to obtain the similarity between the lateral X-ray film and the lateral DRR.

[0064] 4) performing weighted summation of the similarity between the anteroposterior X-ray and the anteroposterior DRR and the similarity between the lateral X-ray and the anteroposterior DRR to obtain a similarity measure;

[0065] Specifically, the similarity between the frontal X-ray film and the frontal DRR and the similarity between the lateral X-ray film and the frontal DRR are weighted and summed to obtain the similarity measurement calculation result,

[0066] 4) Taking the registration parameter as a variable, the similarity measure result is optimized by a particle swarm algorithm so that the similarity measure reaches an extreme value.

[0067] Among them, the number of particles in the particle swarm is set to 10, the maximum number of iterations is set to 100, and when the threshold for exiting the iteration is reached or the maximum number of iterations is reached, the final iteration result is output. The threshold is set to when the difference between two consecutive iteration results is less than or equal to 0.01, it can be judged that the extreme value is reached, and the current six-degree-of-freedom registration parameters are obtained. The six-degree-of-freedom registration parameters are used for subsequent use in the surgical process. The six-degree-of-freedom registration parameters represent the difference between the current posture of the patient's spine and the posture taken before the operation. Using the six-degree-of-freedom registration parameters, the X-ray film taken during the operation can be correctly mapped to the three-dimensional CT data taken before the operation. Specifically, the six-degree-of-freedom parameters represent a posture transformation of the registered three-dimensional object relative to the original position, rx, ry, rz represent its own rotation around the x-axis, y-axis, and z-axis, and tx, ty, tz represent its translation on the three axes.

[0068] Embodiment 2

[0069] In order to evaluate the accuracy of the present invention, the present invention adopts a gold standard discrimination method, in which a DRR image generated by CT data of a known position replaces an X-ray image as a two-dimensional input. By setting the rotation and translation parameters of CT as the gold standard, the generated DRR image is registered with the three-dimensional CT image, and then the registration result is compared with the gold standard to measure the performance of the registration system. Since the X-ray image and the DRR image are the same in the imaging principle, the images of the two are highly similar, and the DRR image generated by the gold standard can accurately evaluate the system performance. The present invention and three registration algorithms whose similarity measures are normalized mutual information, gradient correlation, and normalized mutual correlation are respectively used to run ten times continuously, and then the actual running results of the four algorithms are subtracted from the ideal position for six degrees of freedom to calculate the error of each operation. The statistical results of the average error are shown in Table 1, and the statistical results of the mean square error are shown in Table 2. The ideal six-degree-of-freedom registration parameters are (RX, RY, RZ, TX, TY, TZ) = (5, 0, 3, 10, 0, 20).

[0070] Table 1

[0071]

[0072]

[0073] Table 2

[0074]

[0075] Embodiment 3

[0076] The 2D / 3D medical image registration system of the present invention comprises:

[0077] An acquisition module, used for acquiring an anteroposterior X-ray film and an anteroposterior DRR, and acquiring a lateral X-ray film and a lateral DRR;

[0078] The first calculation module is used to calculate the similarity between the frontal X-ray film and the frontal DRR, and calculate the similarity between the lateral X-ray film and the lateral DRR;

[0079] The second calculation module is used to perform weighted summation on the similarity between the frontal X-ray film and the frontal DRR and the similarity between the lateral X-ray film and the frontal DRR to obtain a similarity measure;

[0080] The optimization module is used to optimize the similarity measurement result by particle swarm algorithm with the registration parameter as a variable, so that the similarity measurement reaches an extreme value, and obtains the registration parameter corresponding to the extreme value of the similarity measurement.

[0081] In this embodiment, the first calculation module includes:

[0082] The third calculation module is used to calculate the normalized mutual information between the anteroposterior X-ray film and the anteroposterior DRR;

[0083] A fourth calculation module is used to calculate the gradient-related information of the anteroposterior X-ray film and the anteroposterior DRR;

[0084] The first weighted summation module is used to perform weighted summation on the normalized mutual information between the orthostatic X-ray film and the orthostatic DRR and the gradient-related information between the orthostatic X-ray film and the orthostatic DRR to obtain the similarity between the orthostatic X-ray film and the orthostatic DRR.

[0085] In this embodiment, the first calculation module includes:

[0086] a fifth calculation module, for calculating the normalized mutual information between the lateral X-ray film and the lateral DRR;

[0087] A sixth calculation module, used to calculate the gradient-related information of the lateral X-ray film and the lateral DRR;

[0088] The second weighted summation module is used to perform weighted summation on the normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR to obtain the similarity between the lateral X-ray film and the lateral DRR.

[0089] The division of modules in the embodiments of the present application is schematic and is only a logical function division. There may be other division methods in actual implementation. In addition, each functional module in each embodiment of the present application may be integrated into a processor, or may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated modules may be implemented in the form of hardware or in the form of software functional modules.

[0090] Embodiment 4

[0091] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the 2D / 3D medical image registration method are implemented, for example, including: obtaining an anteroposterior X-ray film and an anteroposterior DRR, obtaining a lateral X-ray film and a lateral DRR; calculating the similarity between the anteroposterior X-ray film and the anteroposterior DRR, and calculating the similarity between the lateral X-ray film and the lateral DRR; performing weighted summation on the similarity between the anteroposterior X-ray film and the anteroposterior DRR and the similarity between the lateral X-ray film and the anteroposterior DRR to obtain a similarity measure; taking a registration parameter as a variable, optimizing the similarity measure result by a particle swarm algorithm so that the similarity measure reaches an extreme value, and obtaining the registration parameter corresponding to the extreme value of the similarity measure. The memory may include a memory, such as a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk memory, etc. The processor, network interface, and memory are interconnected through an internal bus, and the internal bus may be an industrial standard architecture bus, a peripheral component interconnection standard bus, an extended industrial standard architecture bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs. Specifically, the program may include a program code, and the program code includes computer operation instructions. The memory may include a memory and a non-volatile memory, and provide instructions and data to the processor.

[0092] Embodiment 5

[0093] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the 2D / 3D medical image registration method are implemented, for example, including: obtaining an anteroposterior X-ray film and an anteroposterior DRR, obtaining a lateral X-ray film and a lateral DRR; calculating the similarity between the anteroposterior X-ray film and the anteroposterior DRR, and calculating the similarity between the lateral X-ray film and the lateral DRR; weighted summing the similarity between the anteroposterior X-ray film and the anteroposterior DRR and the similarity between the lateral X-ray film and the anteroposterior DRR to obtain a similarity measure; using a registration parameter as a variable, optimizing the similarity measure result by a particle swarm algorithm so that the similarity measure reaches an extreme value, and obtaining the registration parameter corresponding to the extreme value of the similarity measure. Specifically, the computer-readable storage medium includes, but is not limited to, for example, a volatile memory and / or a non-volatile memory. The volatile memory may include a random access memory (RAM) and / or a cache memory, etc. The non-volatile memory may include a read-only memory (ROM), a hard disk, a flash memory, an optical disk, a magnetic disk, and the like.

[0094] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0095] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A 2D / 3D medical image registration method, characterized in that: include: Get an AP X-ray and AP DRR, get a lateral X-ray and lateral DRR; The similarity between the anteroposterior X-ray and the anteroposterior DRR was calculated, and the similarity between the lateral X-ray and the lateral DRR was calculated; The similarity between the anteroposterior X-ray and the anteroposterior DRR and the similarity between the lateral X-ray and the anteroposterior DRR are weighted summed to obtain the similarity measure; Taking the registration parameter as a variable, the similarity measure result is optimized by a particle swarm algorithm so that the similarity measure reaches an extreme value, and the registration parameter corresponding to the extreme value of the similarity measure is obtained.

2. The 2D / 3D medical image registration method according to claim 1, characterized in that: The process of obtaining the frontal X-ray film and the frontal DRR, and obtaining the lateral X-ray film and the lateral DRR is as follows: Take anteroposterior and lateral X-rays of the patient during the operation; The patient's spine CT data taken before surgery is obtained, and the lateral DRR and anteroposterior DRR of the CT are generated using the ray projection method of the ITK library based on the initial alignment parameters and the patient's spine CT data.

3. The 2D / 3D medical image registration method according to claim 1, characterized in that: Before calculating the similarity between the frontal X-ray film and the frontal DRR and calculating the similarity between the lateral X-ray film and the lateral DRR, the method further includes: The opencv library of python is used to denoise the anteroposterior and lateral X-rays.

4. The 2D / 3D medical image registration method according to claim 1, characterized in that: The process of calculating the similarity between the frontal X-ray film and the frontal DRR is: Calculate the normalized mutual information between the AP X-ray and the AP DRR; Calculate the gradient-related information between the AP X-ray and AP DRR; The normalized mutual information between the orthotopic X-ray film and the orthotopic DRR and the gradient-related information between the orthotopic X-ray film and the orthotopic DRR are weightedly summed to obtain the similarity between the orthotopic X-ray film and the orthotopic DRR.

5. The 2D / 3D medical image registration method according to claim 1, characterized in that: The process of calculating the similarity between the lateral X-ray film and the lateral DRR is as follows: Calculate the normalized mutual information between the lateral X-ray and the lateral DRR; Calculate the gradient-related information between the lateral X-ray and the lateral DRR; The normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR are weightedly summed to obtain the similarity between the lateral X-ray film and the lateral DRR.

6. A 2D / 3D medical image registration system, characterized in that: include: An acquisition module, used for acquiring an anteroposterior X-ray film and an anteroposterior DRR, and acquiring a lateral X-ray film and a lateral DRR; The first calculation module is used to calculate the similarity between the frontal X-ray film and the frontal DRR, and calculate the similarity between the lateral X-ray film and the lateral DRR; The second calculation module is used to perform weighted summation on the similarity between the frontal X-ray film and the frontal DRR and the similarity between the lateral X-ray film and the frontal DRR to obtain a similarity measure; The optimization module is used to optimize the similarity measurement result by particle swarm algorithm with the registration parameter as a variable, so that the similarity measurement reaches an extreme value, and obtains the registration parameter corresponding to the extreme value of the similarity measurement.

7. The 2D / 3D medical image registration system according to claim 6, characterized in that: The first calculation module includes: The third calculation module is used to calculate the normalized mutual information between the anteroposterior X-ray film and the anteroposterior DRR; A fourth calculation module is used to calculate the gradient-related information of the anteroposterior X-ray film and the anteroposterior DRR; The first weighted summation module is used to perform weighted summation on the normalized mutual information between the orthostatic X-ray film and the orthostatic DRR and the gradient-related information between the orthostatic X-ray film and the orthostatic DRR to obtain the similarity between the orthostatic X-ray film and the orthostatic DRR.

8. The 2D / 3D medical image registration system according to claim 6, characterized in that: The first calculation module includes: a fifth calculation module, for calculating the normalized mutual information between the lateral X-ray film and the lateral DRR; A sixth calculation module, for calculating gradient-related information of the lateral X-ray film and the lateral DRR; The second weighted summation module is used to perform weighted summation on the normalized mutual information between the lateral X-ray film and the lateral DRR and the gradient-related information between the lateral X-ray film and the lateral DRR to obtain the similarity between the lateral X-ray film and the lateral DRR.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the 2D / 3D medical image registration method according to any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the 2D / 3D medical image registration method according to any one of claims 1 to 5 are implemented.

Images