Image registration method, apparatus, device, medium, and product for radiotherapy

Abstract

This invention discloses an image registration method, apparatus, device, medium, and product for radiotherapy. The method includes: acquiring a three-dimensional medical image of a target object, a surface image in a predetermined position, and a combination of X-ray images for positioning, the X-ray image combination including X-ray images acquired from a first viewpoint and a second viewpoint respectively; determining a surface contour image corresponding to the three-dimensional medical image; determining a registration matrix for registering the surface contour image to the surface image and a floating registration result of the three-dimensional medical image under the registration matrix; determining projection images of the floating registration result in the first and second viewpoints respectively, under the pixel set corresponding to the X-ray image combination, to obtain a projection image combination; and outputting a registration matrix if the similarity measure between the projection image combination and the X-ray image combination meets predetermined screening conditions. This invention can improve the accuracy of image registration results between three-dimensional medical images and X-ray image combinations.

Description

Technical Field

[0001] This invention relates to the field of image processing technology, and in particular to an image registration method, apparatus, device, medium and product for radiotherapy. Background Technology

[0002] To perform radiotherapy, it is usually necessary to acquire a three-dimensional medical image of the target object (patient) in a predetermined position during the simulation positioning phase. Then, a radiotherapy plan is developed for the target object based on the three-dimensional medical image. During the radiotherapy execution phase, an optical guidance system is used to acquire a surface image of the target object in the predetermined position. The three-dimensional medical image and the surface image are then registered to obtain a registration result, which maps the three-dimensional medical image to the coordinate system of the surface image. Based on the registration result and the radiotherapy plan, the radiotherapy equipment is controlled to perform radiotherapy operations on the target object in the predetermined position.

[0003] In summary, the effectiveness of radiotherapy depends on the accuracy of the registration results mentioned above. However, existing image registration methods cannot guarantee the accuracy of these results. Therefore, radiotherapy performed based on existing image registration methods is unlikely to guarantee good radiotherapy results. Summary of the Invention

[0004] This invention provides an image registration method, apparatus, device, medium, and product for radiotherapy, to address the problem that existing image registration methods often fail to guarantee high accuracy in the image registration results between the combination of three-dimensional medical images acquired during the simulation positioning phase and the X-ray images used for radiotherapy positioning.

[0005] According to one aspect of the present invention, an image registration method for radiotherapy is provided, comprising:

[0006] Acquire three-dimensional medical images for developing a radiotherapy plan for a target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and a combination of X-ray images for positioning. The combination of X-ray images includes X-ray images acquired from a first viewpoint and a second viewpoint, respectively, and the first viewpoint and the second viewpoint are orthogonal.

[0007] Determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix for registering the surface contour image to the surface image, and determine the floating registration result of the three-dimensional medical image under the action of the registration matrix;

[0008] Under the pixel set corresponding to the ray image combination, determine the projected images of the floating registration result under the first viewpoint and the second viewpoint respectively, and obtain the projected image combination;

[0009] Determine the similarity measure between the projected image combination and the ray image combination, and output the registration matrix if the similarity measure meets predetermined screening conditions.

[0010] According to another aspect of the present invention, an image registration apparatus for radiotherapy is provided, comprising:

[0011] The acquisition module is used to acquire three-dimensional medical images for developing a radiotherapy plan for a target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and a combination of X-ray images for positioning. The combination of X-ray images includes X-ray images acquired from a first viewpoint and a second viewpoint, respectively, and the first viewpoint and the second viewpoint are orthogonal.

[0012] The registration module is used to determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix for registering the surface contour image to the surface image, and the floating registration result of the three-dimensional medical image under the action of the registration matrix;

[0013] The projection module is used to determine the projected images of the floating registration result under the first viewing angle and the second viewing angle respectively, under the pixel set corresponding to the ray image combination, so as to obtain the projected image combination;

[0014] The decision module is used to determine the similarity measure between the projected image combination and the ray image combination, and output the registration matrix if the similarity measure meets the predetermined screening conditions.

[0015] According to another aspect of the present invention, a computer device is provided, the computer device comprising:

[0016] At least one processor; and

[0017] A memory communicatively connected to the at least one processor; wherein,

[0018] The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the image registration method for radiotherapy as described in any embodiment of the present invention.

[0019] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the image registration method for radiotherapy as described in any embodiment of the present invention.

[0020] According to another aspect of the present invention, a computer program product is provided, the computer program product comprising a computer program that, when executed by a processor, implements the image registration method for radiotherapy as described in any embodiment.

[0021] The technical solution of this invention, since the combination of ray images can reflect the current position information of the target object, after determining the registration matrix for registering the surface contour image to the surface image and the floating registration result of the three-dimensional medical image under the action of the registration matrix, under the pixel set corresponding to the ray image combination, determines the projection images of the floating registration result in the first and second viewpoints respectively to obtain the projection image combination, and then calculates the similarity measure between the projection image combination and the ray image combination. The introduction of the voxel set helps to improve the similarity measure between the two image combinations, thereby increasing the probability that the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions. Moreover, the registration matrix is ​​output only when the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions, which can ensure the accuracy of the registration result and the effect of radiotherapy performed based on the radiotherapy plan information corresponding to the three-dimensional medical image.

[0022] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0024] Figure 1 A flowchart of an image registration method for radiotherapy provided in an embodiment of the present invention;

[0025] Figure 2A The first ray image and the first projection image provided in the embodiments of the present invention;

[0026] Figure 2B The second ray image and the second projection image provided in the embodiments of the present invention;

[0027] Figure 2C This is a schematic diagram of the superposition result of the first ray image and the first projection image provided in an embodiment of the present invention;

[0028] Figure 2DThis is a schematic diagram of the superposition result of the second ray image and the second projection image provided in an embodiment of the present invention;

[0029] Figure 2E This is a schematic diagram of the image registration result between the first ray image and the first projection image provided in an embodiment of the present invention;

[0030] Figure 2F This is a schematic diagram of the image registration result between the second ray image and the second projection image provided in an embodiment of the present invention;

[0031] Figure 3 This is yet another flowchart of an image registration method for radiotherapy provided according to an embodiment of the present invention;

[0032] Figure 4 This is a schematic diagram of the structure of an image registration device for radiotherapy provided according to an embodiment of the present invention;

[0033] Figure 5 This is a schematic diagram of the structure of a computer device for implementing the image registration method for radiotherapy according to embodiments of the present invention. Detailed Implementation

[0034] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0035] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0036] Figure 1This is a flowchart of an image registration method for radiotherapy provided in an embodiment of the present invention. This embodiment is applicable to situations where a similarity measure between a projection image combination and a radiation image combination is determined based on the pixel set corresponding to the radiation image combination, and whether a registration matrix is ​​output is determined based on whether the similarity measure meets predetermined screening conditions. This method can be executed by an image registration device for radiotherapy, which can be implemented in hardware and / or software, and can be configured in the processor of a computer device. Figure 1 As shown, the method includes:

[0037] S110. Acquire a combination of three-dimensional medical images for developing a radiotherapy plan for the target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and X-ray images for positioning, wherein the X-ray image combination includes X-ray images acquired from a first viewpoint and a second viewpoint, and the first viewpoint and the second viewpoint are orthogonal.

[0038] Three-dimensional medical images used for radiotherapy planning are acquired during the simulation and localization phase. After acquisition, these images are uploaded to a server. Doctors use the radiotherapy planning system to query these images and delineate the target area based on the patient's condition. Physicists then develop a radiotherapy plan targeting the three-dimensional medical image and its corresponding target area, based on the delineated target area and the doctor's radiotherapy orders for that patient. The generated radiotherapy plan is then submitted to the server.

[0039] During the radiotherapy execution phase, the technician guides the target subject to complete the positioning based on the predetermined body position established during the simulation positioning phase. A surface optical guidance system is then used to acquire a surface image of the target subject in this current positioning. The technician then controls a radiographic imaging device to acquire a composite of radiographic images of the target subject in this current positioning. This composite of radiographic images includes a first radiographic image acquired from a first perspective and a second radiographic image acquired from a second perspective, with the first and second perspectives orthogonal. In one embodiment, the first perspective is the AP perspective (anterior-posterior direction), and the second perspective is the LAT perspective (lateral direction). It is understood that the composite of radiographic images best matches the target subject's current positioning and physical condition.

[0040] The predetermined position is a position adapted to the radiotherapy site of the target patient. In this embodiment, the predetermined position can be any industry-standard predetermined position, and no specific limitation is made here.

[0041] Positioning can be understood as replicating a predetermined body position so that the target subject can complete the current radiotherapy in the same predetermined body position used in the simulation positioning phase.

[0042] S120. Determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix used to register the surface contour image to the surface image, and the floating registration result of the three-dimensional medical image under the action of the registration matrix.

[0043] After the surface image and radiographic image are acquired together, the processor automatically retrieves the 3D medical image corresponding to the target object identifier and the corresponding radiotherapy plan information, or, in response to a retrieval operation for the 3D medical image and radiotherapy plan information, retrieves the 3D medical image corresponding to the target object identifier and the corresponding radiotherapy plan information; then, it determines the surface contour image corresponding to the 3D medical image. The surface contour image only includes the body surface information of the target object in the 3D medical image and does not include internal structural information, such as internal organ tissue information.

[0044] The surface contour image corresponding to the three-dimensional medical image in this embodiment can be determined using existing methods, or it can be determined using the following methods:

[0045] Step a1: Determine the energy function for the three-dimensional medical image. The energy function includes a length constraint term, an area constraint term, a first intensity constraint term, and a second intensity constraint term. The first intensity constraint term is used to determine the intensity constraint information corresponding to all pixels in the three-dimensional medical image with pixel values ​​greater than zero. The second intensity constraint term is used to determine the intensity constraint information corresponding to all pixels in the three-dimensional medical image with pixel values ​​less than 1 and greater than zero.

[0046] The energy function is Where C is the surface contour curve of the target object in the 3D medical image, Length(C) is the length constraint term, Area(C) is the area constraint term, E1(C) and E2(C) are the intensity constraints, μ, λ1 and λ2 are constant coefficients. Each constraint term can be specifically expressed as:

[0047]

[0048] in, C1 is the expression for the constrained curve; C2 is the average pixel value of all pixels inside the curve and C1 is the average pixel value of all pixels outside the curve.

[0049] Step a2: Determine the partial differential equation of the energy function by solving for the extreme values ​​of the energy function.

[0050] By using variational methods and gradient descent flow to solve for the extrema of the energy function, the partial differential equation of the energy function can be obtained. Where t is the time step of the evolution curve.

[0051] Step a3: Determine the evolution curve based on the partial differential equation, and use the evolution curve as the surface contour image corresponding to the three-dimensional medical image.

[0052] Solving the above partial differential equation, the evolution curve determined by the iteration cycle corresponding to the iteration termination condition is used as the surface contour image corresponding to the three-dimensional medical image. Existing iteration termination conditions can be used; this embodiment does not impose specific limitations.

[0053] After the surface contour image is determined, image registration is performed between the surface contour image and the surface image to obtain a registration matrix used to register the surface contour image to the surface image. It can be understood that the registration matrix used to register the surface contour image to the surface image represents the transformation relationship between the coordinate system of the imaging device in the simulation positioning phase and the coordinate system of the radiotherapy setup. This registration matrix can also accurately register the 3D medical image to the surface image, and the floating registration result generated by this registration operation has high accuracy. Since the data volume of the surface contour image is much smaller than that of the 3D medical image, compared to directly registering the 3D medical image and the surface image to determine the registration matrix, performing image registration between the surface contour image and the surface image to determine the registration matrix can significantly reduce the amount of data computation in the registration matrix determination process and improve the speed of registration matrix determination.

[0054] In one embodiment, using the surface contour image as the floating image and the surface image as the reference image, a similarity measure between the surface contour image and the surface image is calculated, and this similarity measure is used as the registration similarity measure. Then, the registration matrix is ​​obtained by solving the initial spatial transformation matrix using the least squares method based on the registration similarity measure. This embodiment may optionally use the sum of squared errors as the registration similarity measure.

[0055] The registration matrix can be represented as M = [tx, ty, tz, rx, ry, rz], which can be decomposed into translation transformation T and rotation transformation R, where T = [tx, ty, tz]. The rotation transformation R can be decomposed into the product of rotation transformations in the three orthogonal directions x, y, and z: R = R x ·R y ·R z .in,

[0056]

[0057] By using a registration matrix to transform each voxel Pv in a 3D medical image, the corresponding voxel P in the coordinate system of the radiotherapy equipment can be determined. C Specifically, P C =R·T·P VAfter determining the corresponding voxels of each voxel in the 3D medical image within the coordinate system of the radiotherapy equipment, interpolation, such as trilinear interpolation, can be used to obtain the floating registration image of the 3D medical image in the space where the radiotherapy equipment is located.

[0058] A similarity measure (S) is determined to measure the similarity between the floating registered image and the reference image. The transformation parameters of the similarity measure (S) in six degrees of freedom (translation and rotation in x, y, and z) can be expressed as M = [x_trans, y_trans, z_trans, x_rot, y_rot, z_rot]. A Taylor expansion yields the following results:

[0059]

[0060] Recorded as:

[0061]

[0062] By performing an optimal estimate on the above formula, the iterative formula becomes:

[0063] △t=(A T A) -1 A T b

[0064] When Δt is less than the preset threshold, the iteration terminates, and M, including the transformation parameters, is the registration matrix.

[0065] In one embodiment, before / after / simultaneously determining the registration matrix, an existing image smoothing method, such as Gaussian image smoothing, is used to smooth the 3D medical image to update it. Then, the updated 3D medical image is processed using the registration matrix to obtain a floating registration result. This embodiment improves the image quality of the 3D medical image and the accuracy of the floating registration result by smoothing the 3D medical image.

[0066] Specifically, the desired value is selected based on the size of the three-dimensional medical image. μ Calculate the one-dimensional Gaussian convolution kernel using the standard deviation σ. The three-dimensional medical image is convolved sequentially with the Gaussian convolution kernel along the x, y, and z axes to obtain the updated three-dimensional medical image.

[0067] S130. Under the pixel set corresponding to the ray image combination, determine the projected images of the floating registration result under the first view and the second view respectively, and obtain the projected image combination.

[0068] Take all the pixels in the ray image combination as a pixel set, determine the pixel with the largest x-coordinate value in the pixel set and take that pixel as the first pixel; determine the pixel with the largest y-coordinate value in the pixel set and take that pixel as the second pixel.

[0069] Specifically, the floating registration results are determined as a first initial projected image under a first viewpoint and a second initial projected image under a second viewpoint, respectively, to obtain an initial projected image combination. Pixels with x-coordinate values ​​greater than the x-coordinate value of the first pixel and pixels with y-coordinate values ​​greater than the y-coordinate value of the second pixel are deleted from the initial projected image combination to obtain another projected image combination. Alternatively, non-intersecting pixels from the initial projected image combination and the ray image combination are deleted, while the intersecting pixels are retained, to obtain another projected image combination.

[0070] Regarding the first and second initial projected images: In the coordinate system of the radiotherapy equipment, the spatial coordinates of the radiation source are determined; this is the focal point of the fluoroscopic projection rays and also the origin of all fluoroscopic projection rays. The aforementioned pixel set is taken as the endpoint of the fluoroscopic projection rays. Siddon's projection is performed on the 3D medical image (or a smoothed 3D medical image) to determine the floating registration image voxels traversed by each fluoroscopic projection ray. A weighted average is then calculated based on the intensity of each fluoroscopic projection ray and the length of its intersecting line segment. Specifically, according to... Determine the first initial projection image, based on Determine the second initial projection image. Where α k For AP viewpoint projection ray and I VolumeT The length of the line segment intersecting the k-th voxel, β k For LAT viewpoint projection ray and I VolumeT The length of the line segment intersecting the k-th voxel.

[0071] In one embodiment, each ray image in the ray image combination is smoothed to obtain an updated ray image combination; under the pixel set corresponding to the updated ray image combination, the projected images of the floating registration result under the first viewpoint and the second viewpoint are determined to obtain a projected image combination. By smoothing each ray image in the ray image combination, the image quality of each ray image in the ray image combination can be improved, such as the smoothness of the ray images, thereby improving the accuracy of the pixel set corresponding to the ray image combination.

[0072] Specifically, μ and standard deviation σ are selected based on the size of each ray image in the ray image combination, and a one-dimensional Gaussian convolution kernel is calculated. The first ray image and the second ray image are convolved with the Gaussian convolution kernel sequentially on the x, y, and z axes to obtain the updated first ray image and the second ray image.

[0073] S140. Determine the similarity measure between the combination of projected images and the combination of ray images, and output the registration matrix if the similarity measure meets the predetermined screening conditions.

[0074] A similarity measure is a quantitative metric that characterizes or describes the degree of similarity between matching entities. It is typically calculated using a cost function.

[0075] In this embodiment, the similarity measure is used to describe the overall similarity between the ray image combination and the projection image combination. The similarity measure is the mean of the similarity between the corresponding images in the ray image combination and the projection image combination.

[0076] Specifically, a first similarity measure between the ray image and the projected image from a first perspective and a second similarity measure between the ray image and the projected image from a second perspective are determined, and the mean between the first similarity measure and the second similarity measure is used as the similarity measure between the combination of ray images and the combination of projected images.

[0077] Furthermore, the ray image combination includes a first ray image corresponding to the first viewpoint and a second ray image corresponding to the second viewpoint, and the projection image combination includes a first projection image corresponding to the first viewpoint and a second projection image corresponding to the second viewpoint. After the ray image combination and the projection image combination are determined, the sum of squared errors between corresponding pixels of the first projection image and the first ray image is determined, and all sums of squared errors between the first projection image and the first ray image are used as a first similarity measure; the sum of squared errors between corresponding pixels of the second projection image and the second ray image is determined, and all sums of squared errors between the second projection image and the second ray image are used as a second similarity measure; the mean of the first similarity measure and the second similarity measure is used as the similarity measure between the projection image combination and the ray image combination.

[0078] In summary, the first similarity measure reflects the degree of similarity between the first projected image and the first ray image, and the lower the first similarity measure, the higher the similarity between the two images. Similarly, the second similarity measure reflects the degree of similarity between the second projected image and the second ray image, and the lower the second similarity measure, the higher the similarity between the two images. Likewise, the lower the similarity measure, the higher the similarity between the combination of projected images and the combination of ray images.

[0079] The predetermined screening criteria are those that enable the cost function to reach an extreme value, such as a minimum value.

[0080] After the registration matrix is ​​determined, the processor determines the image registration result between the three-dimensional medical image (or the smoothed three-dimensional medical image) and the surface image based on the registration matrix; it reads the radiotherapy plan information corresponding to the three-dimensional medical image, and generates control information for controlling the radiotherapy equipment to perform radiotherapy based on the image registration result and the radiotherapy plan information, so that the radiotherapy equipment can perform radiotherapy on the target object according to the control information.

[0081] In this embodiment, if the similarity measure is the minimum value of the cost function, it is determined that the similarity measure meets the predetermined screening conditions, and the surface contour image corresponding to the current combination of projected images is the required target surface contour image. Therefore, the registration matrix is ​​output.

[0082] Figure 2A The green image on the left is the first ray image under the AP view (first view) provided in the embodiment of the present invention. Figure 2A The red image on the right is the first projected image from the AP perspective (first perspective) provided in this embodiment of the invention; Figure 2B The green image on the left is the second ray image under the LAP perspective (second perspective) provided in the embodiment of the present invention. Figure 2B The red image on the right is the second projection image under the LAP perspective (second perspective) provided in the embodiment of the present invention; Figure 2C This is a schematic diagram showing the superposition result of the first ray image and the first projection image provided in an embodiment of the present invention. Figure 2D This is a schematic diagram of the superposition result of the second ray image and the second projection image provided in an embodiment of the present invention; Figure 2E This is a schematic diagram of the image registration result between the first ray image and the first projection image provided in an embodiment of the present invention; Figure 2F This is a schematic diagram illustrating the image registration result between the second ray image and the second projection image provided in an embodiment of the present invention. It is obvious that, compared to... Figure 2C and Figure 2D , Figure 2E and Figure 2F The image registration results obtained in this invention exhibit high accuracy. Therefore, the registration matrix determined in this embodiment can accurately complete the registration of the combination of three-dimensional medical images and X-ray images, resulting in highly accurate image registration results. Highly accurate image registration results can ensure the effectiveness of radiotherapy.

[0083] The technical solution provided by this invention, since the combination of ray images can reflect the current positional information of the target object, after determining the registration matrix for registering the surface contour image to the surface image and the floating registration result of the three-dimensional medical image under the action of the registration matrix, under the pixel set corresponding to the ray image combination, the projection images of the floating registration result under the first view and the second view are determined to obtain the projection image combination. Then, the similarity measure between the projection image combination and the ray image combination is calculated. The introduction of the voxel set helps to improve the similarity measure between the two image combinations, thereby increasing the probability that the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions. Moreover, the registration matrix is ​​output only when the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions, which can ensure the accuracy of the registration result and the effect of radiotherapy performed based on the radiotherapy plan information corresponding to the three-dimensional medical image.

[0084] Figure 3 This is another flowchart of the image registration method for radiotherapy provided in this embodiment of the invention. This embodiment refines the steps for determining the combination of projection images. Figure 3 As shown, the method includes:

[0085] S210. Acquire a combination of three-dimensional medical images for developing a radiotherapy plan for the target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and X-ray images for positioning, wherein the X-ray image combination includes X-ray images acquired from a first viewpoint and a second viewpoint, and the first viewpoint and the second viewpoint are orthogonal.

[0086] S220. Determine the surface contour image corresponding to the three-dimensional medical image, and determine the registration matrix used to register the surface contour image to the surface image and the floating registration result of the three-dimensional medical image under the action of the registration matrix.

[0087] S230. Under the pixel set corresponding to the ray image combination, determine the projected images of the floating registration result under the first viewpoint and the second viewpoint respectively, and obtain the projected image combination.

[0088] S2401. Determine whether the similarity measure between the combination of projected images and the combination of ray images meets the predetermined screening criteria.

[0089] S2402. If so, output the registration matrix.

[0090] S2403. If not, determine the contribution of each pixel in the pixel set to the similarity measure, and select pixels that meet the contribution condition from the pixel set to update the pixel set.

[0091] In one embodiment, the contribution of each pixel in the pixel set to the similarity measure is determined based on image gradient information, as follows:

[0092] Step b1: Determine the target residual vector of the projected image combination relative to the ray image combination. The target residual vector includes a first residual vector between the projected image and the ray image corresponding to the first viewpoint, and a second residual vector between the projected image and the ray image corresponding to the second viewpoint.

[0093] The first residual vector of the projected image from the first-view perspective compared to the ray image from the first-view perspective can be expressed as: Among them, I DRR-AP (n) represents the grayscale value of the pixel identified as n in the projected image from the first-view perspective. XG-AP (n) represents the gray value of the pixel identified as n in the ray image from the first-view perspective.

[0094] The second residual vector of the projected image from the second viewpoint compared to the ray image from the second viewpoint can be expressed as: Among them, I DRR-LAT (n) represents the grayscale value of the pixel identified as n in the projected image from the second perspective. XG-LAT (n) represents the gray value of the pixel identified as n in the ray image from the second perspective, b AP With b LAT Each includes n elements.

[0095] The target residual vector can be expressed as:

[0096] Step b2: Determine the target matrix based on the target residual vector. The target matrix includes the derivatives of each element of the target residual vector in each predetermined degree of freedom direction.

[0097] The target matrix includes the derivatives of each element in b with respect to the six degrees of freedom, and can be specifically expressed as follows: Among them, b n This represents the element identified as n in b. These six degrees of freedom include translational degrees of freedom and rotational degrees of freedom along each coordinate axis. Specifically, the six degrees of freedom include the translational degree of freedom tx along the x-axis, the translational degree of freedom ty along the y-axis, the translational degree of freedom tz along the z-axis, the rotational degree of freedom rx along the x-axis, the rotational degree of freedom ty along the y-axis, and the rotational degree of freedom rz along the z-axis.

[0098] Step b3: Take the concatenation result of the target matrix and the target residual vector as the concatenation matrix, and determine the self-multiplication result of the concatenation matrix.

[0099] The concatenated matrix can be represented as Alpha = [A b]. The concatenation result is the product of the transpose of the concatenated matrix and the concatenated matrix itself, specifically expressed as: Z = Alpha T Alpha.

[0100] Step b4: Determine the target vector including the inner product of each row vector of the concatenation matrix, and use the difference between the target vector and the self-multiplication result as the contribution vector, which includes the contribution of each pixel in the pixel set to the similarity measure.

[0101] After the contribution vector is determined, all elements in the contribution vector are sorted. Taking descending order as an example, the top N% of elements in the sorted result are combined as the updated pixel set. Here, N% is the contribution percentage threshold, and N can be selected as 80, 85, 90, etc. This achieves the goal of preserving image information to the greatest extent while reducing the computational load of similarity measurement and improving the speed of determining similarity measurement and registration matrix.

[0102] Since both the target residual vector and the target matrix correspond to the projected image combination, and the projected image combination is associated with the pixel set, the target residual vector and the target matrix are updated synchronously after the pixel set is updated. Specifically, the elements in the target residual vector that do not belong to the updated pixel set are deleted, and only the elements corresponding to the pixels in the updated pixel set are retained; similarly, the elements in the target matrix that do not belong to the updated pixel set are deleted, and only the elements corresponding to the pixels in the updated pixel set are retained.

[0103] S2404. Under the updated pixel set, the derivatives of the projected image combination in each degree of freedom and the target residual vector of the projected image combination relative to the ray image combination are input into the step size update expression to obtain the update step size of the registration matrix. The step size update expression is determined by solving an optimization problem using the least squares method. The optimization problem is the minimization problem of the target residual vector, where the target residual vector is the residual vector of the projected image combination relative to the ray image combination.

[0104] The step size update expression can be represented as ΔM = (A T A) -1 A T b, the calculation result can be expressed as: ΔM=[Δtx,Δty,Δtz,Δrx,Δry,Δrz].

[0105] The step size update expression is determined by solving the optimization problem using the least squares method. The optimization problem can be expressed as:

[0106] argmin||At-b|| 2 2

[0107] This embodiment utilizes the chain rule and vector orthogonality to allow the use of the same gradient matrix in each iteration, eliminating the need to recalculate the image's gradient information after each iteration. This reduces the computational load of the algorithm and improves its running speed.

[0108] S2405. Update the registration matrix based on the registration matrix and the update step size of the registration matrix, and perform image transformation on the three-dimensional medical image based on the updated registration matrix to obtain the updated floating registration result.

[0109] Since the floating registration result is determined based on the registration matrix, after the update step size of the registration matrix is ​​determined, the sum of the current registration matrix and the update step size is used as the updated registration matrix. Then, based on the updated registration matrix, image transformation is performed on the 3D medical image or the smoothed 3D medical image to obtain the updated floating registration result.

[0110] S2406. Return the projected images of the determined floating registration results in the first and second perspectives respectively, and obtain the combination of projected images until the current similarity measure meets the predetermined screening conditions or the number of iterations reaches the predetermined number of conditions, and output the current registration matrix.

[0111] Once the updated floating registration result is determined, return to S230 until the similarity measure corresponding to the current registration matrix meets the predetermined screening conditions, or the number of iterations reaches the set number condition, then determine that the iteration has ended and output the current registration matrix.

[0112] The technical solution provided by the embodiments of the present invention updates the pixel set in any iteration by selecting pixels that meet the pose transformation contribution condition from the current pixel set, and updates the projection image combination by updating the pixel set, thereby updating the similarity measure, improving the accuracy of similarity measure update and the accuracy of selecting the registration matrix based on the similarity measure.

[0113] Figure 4 This is a schematic diagram of the image registration device for radiotherapy provided in an embodiment of the present invention. Figure 4 As shown, the device includes:

[0114] The acquisition module 31 is used to acquire three-dimensional medical images for developing a radiotherapy plan for a target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and a combination of X-ray images for positioning. The combination of X-ray images includes X-ray images acquired from a first viewpoint and a second viewpoint, respectively, and the first viewpoint and the second viewpoint are orthogonal.

[0115] Registration module 32 is used to determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix for registering the surface contour image to the surface image, and the floating registration result of the three-dimensional medical image under the action of the registration matrix;

[0116] Projection module 33 is used to determine the projected images of the floating registration result under the first viewing angle and the second viewing angle respectively under the pixel set corresponding to the ray image combination, so as to obtain the projected image combination;

[0117] The decision module 34 is used to determine the similarity measure between the projected image combination and the ray image combination, and output the registration matrix if the similarity measure meets the predetermined screening conditions.

[0118] In one embodiment, the registration module 32 completes the floating registration result of the surface contour image under the action of the registration matrix through the registration unit, which is used for:

[0119] The three-dimensional medical image is smoothed to obtain an updated three-dimensional medical image;

[0120] The updated 3D medical image is spatially transformed using the registration matrix to obtain a floating registration result.

[0121] The step of determining the projected images of the floating registration result under the first and second viewing angles respectively, under the pixel set corresponding to the ray image combination, to obtain the projected image combination includes:

[0122] Smooth each ray image in the ray image combination to obtain the updated ray image combination;

[0123] Under the pixel set corresponding to the updated ray image combination, the projected images of the floating registration result under the first viewpoint and the second viewpoint are determined to obtain the projected image combination.

[0124] In one embodiment, the registration module is used for:

[0125] Using the surface contour image as a floating image and the surface image as a reference image, the similarity measure between the surface contour image and the surface image is calculated, and this similarity measure is used as the registration similarity measure.

[0126] Using the least squares method, the initial spatial transformation matrix is ​​solved based on the registration similarity measure to obtain the registration matrix used to register the surface contour image to the surface image.

[0127] In one embodiment, the decision module 34 is further configured to:

[0128] A pixel selection unit is used to determine the contribution of each pixel in the pixel set to the similarity measure if the similarity measure does not meet the predetermined filtering conditions, and to select pixels that meet the contribution conditions from the pixel set to update the pixel set.

[0129] The step size update unit is used to input the derivatives of the projected image combination in each degree of freedom and the residual vector of the projected image combination relative to the ray image combination into the step size update expression under the updated pixel set to obtain the update step size of the registration matrix. The step size update expression is determined by solving an optimization problem using the least squares method. The optimization problem is a target residual vector minimization problem, and the target residual vector is the residual vector of the projected image combination relative to the ray image combination.

[0130] The registration matrix update unit is used to update the registration matrix based on the registration matrix and the update step size of the registration matrix, and to perform image transformation on the three-dimensional medical image based on the updated registration matrix to obtain the updated floating registration result.

[0131] The unit is used to return the projected images of the determined floating registration result under the first and second viewpoints, and to obtain the combination of the projected images, until the current similarity measure meets the predetermined screening conditions or the number of iterations reaches the predetermined number, and then outputs the current registration matrix.

[0132] In one embodiment, the pixel selection unit determines the contribution of each pixel in the pixel set to the similarity measure through a contribution subunit, which is specifically used for:

[0133] Determine the target residual vector of the projected image combination relative to the ray image combination, the target residual vector including a first residual vector between the projected image and the ray image corresponding to the first viewpoint, and a second residual vector between the projected image and the ray image corresponding to the second viewpoint;

[0134] A target matrix is ​​determined based on the target residual vector, and the target matrix includes the derivatives of each element of the target residual vector in each predetermined degree of freedom direction;

[0135] The concatenation result of the target matrix and the target residual vector is used as the concatenation matrix, and the self-multiplication result of the concatenation matrix is ​​determined.

[0136] A target vector is determined, comprising the inner product of each row vector of the concatenated matrix, and the difference between the target vector and the self-multiplication result is used as a contribution vector, wherein the contribution vector comprises the contribution of each pixel in the pixel set to the similarity measure.

[0137] In one embodiment, the ray image combination includes a first ray image corresponding to the first viewpoint and a second ray image corresponding to the second viewpoint, and the projection image combination includes a first projection image corresponding to the first viewpoint and a second projection image corresponding to the second viewpoint. The similarity measurement module is used for:

[0138] Determine the sum of squared errors between corresponding pixels in the first projected image and the first ray image, and use all the sums of squared errors between the first projected image and the first ray image as the first similarity measure;

[0139] Determine the sum of squared errors between corresponding pixels in the second projected image and the second ray image, and use all the sums of squared errors between the second projected image and the second ray image as the second similarity measure;

[0140] The average of the first similarity measure and the second similarity measure is used as the similarity measure between the projected image combination and the ray image combination.

[0141] In one embodiment, the registration matrix module determines the surface contour image corresponding to the three-dimensional medical image through surface contour units, which are specifically used for:

[0142] An energy function is determined for the three-dimensional medical image. The energy function includes a length constraint term, an area constraint term, a first intensity constraint term, and a second intensity constraint term. The first intensity constraint term is used to determine the intensity constraint information corresponding to all pixels in the three-dimensional medical image whose pixel values ​​are greater than zero. The second intensity constraint term is used to determine the intensity constraint information corresponding to all pixels in the three-dimensional medical image whose pixel values ​​are less than 1 and greater than zero.

[0143] The partial differential equation of the energy function is determined by solving for the extreme values ​​of the energy function.

[0144] The evolution curve is determined based on the partial differential equation, and the evolution curve is used as the surface contour image corresponding to the three-dimensional medical image.

[0145] The technical solution provided by this invention, since the combination of ray images can reflect the current positional information of the target object, after determining the registration matrix for registering the surface contour image to the surface image and the floating registration result of the three-dimensional medical image under the action of the registration matrix, under the pixel set corresponding to the ray image combination, the projection images of the floating registration result under the first view and the second view are determined to obtain the projection image combination. Then, the similarity measure between the projection image combination and the ray image combination is calculated. The introduction of the voxel set helps to improve the similarity measure between the two image combinations, thereby increasing the probability that the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions. Moreover, the registration matrix is ​​output only when the similarity measure between the projection image combination and the ray image combination meets the predetermined screening conditions, which can ensure the accuracy of the registration result and the effect of radiotherapy performed based on the radiotherapy plan information corresponding to the three-dimensional medical image.

[0146] The image registration device for radiotherapy provided in the embodiments of the present invention can execute the image registration method for radiotherapy provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the method.

[0147] Figure 5 A schematic diagram of a computer device 10 that can be used to implement embodiments of the present invention is shown. The computer device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0148] like Figure 5 As shown, the computer device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer programs stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 may also store various programs and data required for the operation of the computer device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0149] Multiple components in computer device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of monitors, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows computer device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0150] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as image registration methods for radiotherapy.

[0151] In some embodiments, the image registration method for radiotherapy may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on computer device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the image registration method for radiotherapy described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the image registration method for radiotherapy by any other suitable means (e.g., by means of firmware).

[0152] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0153] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0154] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0155] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer device having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0156] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0157] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0158] This invention also provides a computer program product, including a computer program that, when executed by a processor, implements the image registration apparatus for radiotherapy as provided in any embodiment of this application.

[0159] In implementing the computer program product, computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages ​​such as Java, Smalltalk, and C++, as well as conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0160] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0161] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An image registration method for radiotherapy, characterized in that, include: Acquire three-dimensional medical images for developing a radiotherapy plan for a target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and a combination of X-ray images based on the positioning results during the radiotherapy execution phase. The combination of X-ray images includes X-ray images acquired from a first viewpoint and a second viewpoint, respectively, and the first viewpoint and the second viewpoint are orthogonal. Determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix for registering the surface contour image to the surface image, and determine the floating registration result of the three-dimensional medical image under the action of the registration matrix; Under the pixel set corresponding to the ray image combination, determine the projection images of the floating registration result under the first viewpoint and the second viewpoint respectively to obtain the projection image combination, including: determining the first initial projection image of the floating registration result under the first viewpoint and the second initial projection image under the second viewpoint respectively to obtain the initial projection image combination; deleting the non-intersecting pixels of the initial projection image combination and the pixel set from the initial projection image combination to obtain the projection image combination. Determine the similarity measure between the projected image combination and the ray image combination, and output the registration matrix if the similarity measure meets predetermined screening conditions.

2. The method according to claim 1, characterized in that, The floating registration result of the three-dimensional medical image under the action of the registration matrix is ​​determined by the following steps: The three-dimensional medical image is smoothed to obtain an updated three-dimensional medical image; The updated 3D medical image is spatially transformed using the registration matrix to obtain a floating registration result. The step of determining the projected images of the floating registration result under the first and second viewing angles respectively, under the pixel set corresponding to the ray image combination, to obtain the projected image combination includes: Smooth each ray image in the ray image combination to obtain the updated ray image combination; Under the pixel set corresponding to the updated ray image combination, the projected images of the floating registration result under the first viewpoint and the second viewpoint are determined to obtain the projected image combination.

3. The method according to claim 1 or 2, characterized in that, Determining the registration matrix used to register the surface contour image to the surface image includes: Using the surface contour image as a floating image and the surface image as a reference image, the similarity measure between the surface contour image and the surface image is calculated, and this similarity measure is used as the registration similarity measure. Using the least squares method, the initial spatial transformation matrix is ​​solved based on the registration similarity measure to obtain the registration matrix used to register the surface contour image to the surface image.

4. The method according to claim 1, characterized in that, After determining the similarity measure between the combination of ray images and the combination of projection images, the method further includes: If the similarity measure does not meet the predetermined filtering conditions, the contribution of each pixel in the pixel set to the similarity measure is determined, and pixels that meet the contribution conditions are selected from the pixel set to update the pixel set. Under the updated pixel set, the derivatives of the projected image combination in each degree of freedom and the residual vector of the projected image combination compared with the ray image combination are input into the step size update expression to obtain the update step size of the registration matrix. The step size update expression is determined by solving an optimization problem, which is the objective residual vector minimization problem. The registration matrix is ​​updated based on the registration matrix and the update step size of the registration matrix, and the three-dimensional medical image is transformed based on the updated registration matrix to obtain the updated floating registration result. The process of determining the projected images of the floating registration result under the first and second viewpoints and obtaining the combined projected images continues until the current similarity measure meets the predetermined screening conditions or the number of iterations reaches the predetermined number of conditions, and the current registration matrix is ​​output. Wherein, under the updated pixel set, the step size update expression is obtained by inputting the derivatives of the projected image combination at each degree of freedom and the residual vector of the projected image combination compared to the ray image combination into the step size update expression, including: Determine the target residual vector of the projected image combination relative to the ray image combination; Delete the elements in the target residual vector that do not belong to the updated pixel set, and update the target residual vector. The derivatives of the projected images at each degree of freedom and the updated residual vector are input into the step size update expression to obtain the update step size of the registration matrix.

5. The method according to claim 4, characterized in that, Determining the contribution of each pixel in the pixel set to the similarity measure includes: Determine the target residual vector of the projected image combination relative to the ray image combination, the target residual vector including a first residual vector between the projected image and the ray image corresponding to the first viewpoint, and a second residual vector between the projected image and the ray image corresponding to the second viewpoint; A target matrix is ​​determined based on the target residual vector, and the target matrix includes the derivatives of each element of the target residual vector in each predetermined degree of freedom direction; The concatenation result of the target matrix and the target residual vector is taken as the concatenation matrix, and the self-multiplication result of the concatenation matrix is ​​determined. A target vector is determined, comprising the inner product of each row vector of the concatenated matrix, and the difference between the target vector and the self-multiplication result is used as a contribution vector, wherein the contribution vector comprises the contribution of each pixel in the pixel set to the similarity measure.

6. The method according to claim 1, characterized in that, The ray image combination includes a first ray image corresponding to the first viewpoint and a second ray image corresponding to the second viewpoint; the projection image combination includes a first projection image corresponding to the first viewpoint and a second projection image corresponding to the second viewpoint; determining the similarity measure between the projection image combination and the ray image combination includes: Determine the sum of squared errors between corresponding pixels in the first projected image and the first ray image, and use all the sums of squared errors between the first projected image and the first ray image as the first similarity measure; Determine the sum of squared errors between corresponding pixels in the second projected image and the second ray image, and use all the sums of squared errors between the second projected image and the second ray image as the second similarity measure; The average of the first similarity measure and the second similarity measure is used as the similarity measure between the projected image combination and the ray image combination.

7. An image registration device for radiotherapy, characterized in that, include: The acquisition module is used to acquire three-dimensional medical images for developing a radiotherapy plan for a target object, surface images of the target object in a predetermined position during the radiotherapy execution phase, and a combination of X-ray images based on the positioning results during the radiotherapy execution phase. The combination of X-ray images includes X-ray images acquired from a first viewpoint and a second viewpoint, respectively, and the first viewpoint and the second viewpoint are orthogonal. The registration matrix module is used to determine the surface contour image corresponding to the three-dimensional medical image, determine the registration matrix used to register the surface contour image to the surface image, and the floating registration result of the three-dimensional medical image under the action of the registration matrix; A projection module is used to determine, under the pixel set corresponding to the ray image combination, the projected images of the floating registration result under the first viewing angle and the second viewing angle respectively, to obtain a projected image combination, including: determining a first initial projected image of the floating registration result under the first viewing angle and a second initial projected image under the second viewing angle respectively, to obtain an initial projected image combination; deleting non-intersecting pixels between the initial projected image combination and the pixel set from the initial projected image combination, to obtain a projected image combination; The decision module is used to determine the similarity measure between the projected image combination and the ray image combination, and output the registration matrix if the similarity measure meets the predetermined screening conditions.

8. A computer device, characterized in that, The computer device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, which enables the at least one processor to perform the image registration method for radiotherapy as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the image registration method for radiotherapy as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the image registration method for radiotherapy according to any one of claims 1-6.

Images