Deformation registration method and device, storage medium and electronic equipment

Abstract

The invention discloses a deformation registration method and device, a storage medium and electronic equipment, and relates to the field of medical science and technology and the field of image processing. The method comprises the following steps: acquiring reference data and floating data; according to the organ type, data separation is carried out on the reference data and the floating data, and the data separation is used for separating the reference data and the floating data to obtain image data of each independent organ; performing deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ; and fusing the deformation field corresponding to each independent organ into a target deformation field, and performing deformation registration on the floating data through the target deformation field. The technical problem that in the prior art, when a registration algorithm is adopted for dose accumulation in radiotherapy, due to the fact that alignment is redundant or insufficient in the registration process of a local organ tissue structure, dose accumulation in an organ boundary area is inaccurate is solved.

Description

Technical Field

[0001] The present application relates to the fields of medical technology and image processing, and specifically to a deformation registration method, device, storage medium and electronic device. Background Art

[0002] In radiotherapy, algorithms are often used to convert dose distributions across treatment phases into a unified coordinate system for dose accumulation. This helps physicians assess the patient's actual total dose and dose distribution, optimizing clinical decision-making. Commonly used dose accumulation methods include linear DVH parameter accumulation, rigid body registration, deformable registration, and biological dose accumulation.

[0003] Among them, deformable registration can handle changes in tissue structure caused by non-rigid motion and reduce dose accumulation errors caused by anatomical deformation, making it a key technology for achieving high-precision dose assessment. However, during the multi-organ delineation and fusion registration process, adjacent organs are constrained by each other, and the deformation fields calculated from different data also vary significantly, making the fusion process more difficult.

[0004] Even if multiple evaluation metrics are introduced into the registration function and local tissue structures are adjusted under the premise of global optimization, although the alignment accuracy of local structures can be improved to a certain extent, it is still essentially a global registration, and the core is still the pursuit of overall optimization. Improvements made only by optimizing the registration function will still result in insufficient local registration accuracy. In general, existing deformable registration methods are mostly based on mathematical models, which drive the global deformation of the image by designing registration functions. They often have redundancy or insufficiency in the alignment of local anatomical structure contours, which can easily lead to inaccurate dose mapping in the boundary areas of tissue structures, thereby affecting the accuracy and reliability of dose accumulation.

[0005] To address the above-mentioned problems, no effective solutions have been proposed so far. Summary of the Invention

[0006] The present application provides a deformable registration method, device, storage medium and electronic device to at least solve the technical problem in the prior art of inaccurate dose accumulation in organ boundary areas due to redundancy or insufficiency in the registration process of local organ tissue structures when using a registration algorithm for dose accumulation in radiotherapy.

[0007] According to one aspect of the present application, a deformable registration method is provided, comprising: acquiring reference data and floating data, wherein the reference data comprises: a reference image, a delineated image of the reference image, and a dose map corresponding to the reference image; and the floating data comprises: a floating image, a delineated image of the floating image, and a dose map corresponding to the floating image; performing data separation on the reference data and the floating data according to organ type, wherein the data separation is used to separate the image data of each independent organ from the reference data and the floating data; performing deformable registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ; fusing the deformation field corresponding to each independent organ into a target deformation field, and performing deformable registration on the floating data using the target deformation field.

[0008] Optionally, deformation registration is performed on the floating data through the target deformation field, including: determining a rigid transformation matrix based on the floating image; performing a rigid transformation on the floating data based on the rigid transformation matrix to obtain intermediate floating data, wherein the intermediate floating data includes the floating image after the rigid transformation, the outline image and the dose map; and deformation registration is performed on the intermediate floating data through the target deformation field.

[0009] Optionally, the reference data and floating data are separated according to the organ type, including: separating each outlined area separately according to the outline information of each organ in the image to obtain N independent outlined reference data of the same size as the reference data and N independent outlined floating data of the same size as the floating data; dividing the reference image into N independent organ reference images according to the N independent outlined reference data corresponding to the reference data, and using the image area of ​​the reference image other than the N independent organ reference images as the background reference image; dividing the floating image into N independent organ floating images according to the N independent outlined floating data corresponding to the floating data, and using the image area of ​​the floating image other than the N independent organ floating images as the background floating image.

[0010] Optionally, deformation registration is performed on the image data of each independent organ to obtain a deformation field corresponding to each independent organ, including: constructing a first function, a second function and a third function, wherein the first function is used to correct the contour alignment error between the floating data and the reference data based on each organ; the second function is used to correct the contour registration redundancy between the floating data and the reference data based on each organ, and the contour registration redundancy characterizes that the contour of the organ after registration exceeds the boundary range corresponding to the organ in the reference image; the third function is used to constrain the degree of deformation of each organ during the registration process; a weighted sum calculation is performed on the first function, the second function and the third function to obtain a target function; and deformation registration is performed on the image data of each independent organ according to the target function to obtain a deformation field corresponding to each independent organ.

[0011] Optionally, constructing a first function, a second function and a third function includes: constructing a first function and a second function based on the independent outline reference data and the independent outline floating data corresponding to the i-th organ, wherein i is a positive integer, and the first function is used to determine the content overlap between the independent outline reference data and the independent outline floating data corresponding to the i-th organ; the second function is used to determine the contour redundancy of the independent outline reference data corresponding to the i-th organ relative to the independent outline floating data corresponding to the i-th organ, and determining the contour redundancy of the independent outline floating data corresponding to the i-th organ relative to the independent outline reference data corresponding to the i-th organ; constructing a third function based on the independent organ floating image and the independent organ reference image corresponding to the i-th organ, wherein the third function is used to determine the pixel overlap between the independent organ floating image and the independent organ reference image corresponding to the i-th organ.

[0012] Optionally, the deformation field corresponding to each independent organ is fused into a target deformation field, including: adding and summing the deformation fields corresponding to each independent organ to obtain the organ deformation field; obtaining the deformation field of the image background area between the reference data and the floating data; obtaining the deformation field of the overlapping area between the image background area and the area where the organ is located; and summing the deformation field of the organ, the deformation field of the image background area, and the deformation field of the overlapping area to obtain the target deformation field.

[0013] Optionally, obtaining a deformation field between the reference data and the floating data regarding the image background area includes: constructing a fourth function based on the background floating image and the background reference image, wherein the fourth function is used to determine the deformation alignment between the background floating image and the background reference image; and performing deformation registration on the background floating image according to the fourth function to obtain the deformation field of the image background area.

[0014] Optionally, obtaining the deformation field of the overlapping area between the image background area and the area where the organ is located includes: determining a mask corresponding to the overlapping area between the image background area and the area where the organ is located based on the intersection between the deformation field of the organ and the deformation field of the image background area; extracting first deformation field data corresponding to the overlapping area from the deformation field of the organ according to the mask; extracting second deformation field data corresponding to the overlapping area from the deformation field of the image background area according to the mask; and performing weighted sum calculation on the first deformation field data and the second deformation field data to obtain the deformation field of the overlapping area.

[0015] According to another aspect of the present application, a deformation registration device is also provided, including: an acquisition unit, used to acquire reference data and floating data, wherein the reference data includes: a reference image, an outline image of the reference image, and a dose map corresponding to the reference image; the floating data includes: a floating image, an outline image of the floating image, and a dose map corresponding to the floating image; a data separation unit, used to perform data separation on the reference data and the floating data respectively according to the organ type, wherein the data separation is used to separate the image data of each independent organ from the reference data and the floating data; a first registration unit, used to perform deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ; and a second registration unit, used to fuse the deformation field corresponding to each independent organ into a target deformation field, and perform deformation registration on the floating data through the target deformation field.

[0016] According to another aspect of the present application, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium. When the computer program runs, the device where the computer-readable storage medium is located executes the above-mentioned deformation registration method.

[0017] According to another aspect of the present application, an electronic device is also provided, wherein the electronic device includes one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors execute the above-mentioned deformation registration method.

[0018] In this application, reference data and floating data are first acquired, wherein the reference data includes: a reference image, an outline image of the reference image, and a dose map corresponding to the reference image; and the floating data includes: a floating image, an outline image of the floating image, and a dose map corresponding to the floating image. Next, data separation is performed on the reference data and floating data based on organ type, wherein data separation is used to separate the image data of each independent organ from the reference data and floating data. Subsequently, deformation registration is performed on the image data of each independent organ to obtain a deformation field corresponding to each independent organ. The deformation fields corresponding to each independent organ are then fused into a target deformation field, and deformation registration is performed on the floating data using the target deformation field.

[0019] As can be seen from the above, this application first separates the reference data and floating data according to the organ type, thereby obtaining the image data of each independent organ in the reference data and floating data. On this basis, this application can perform deformation registration on each independent organ to obtain the deformation field corresponding to each independent organ. Subsequently, a specific fusion strategy is used to fuse the deformation fields corresponding to each independent organ into an overall deformation field. Finally, the fused overall deformation field is used to generate the overall data after deformation registration.

[0020] It's important to note that existing registration techniques typically use all organ information and the entire image as input to perform global registration. These techniques focus on achieving the optimal solution for global deformation, ignoring the differences in deformation characteristics between different organs. Because each organ has different deformation requirements, a unified global deformation often fails to achieve optimal alignment for all organs simultaneously. Furthermore, mutual constraints between different organs further affect the accuracy of local registration.

[0021] This application replaces the traditional global registration approach by splitting the overall delineation and image information into multiple independent delineation and image information corresponding to different organs based on organ type. By performing independent registration for each organ, interference between different organs is avoided, effectively resolving the problem of redundant or insufficient alignment of local organ tissue structures during the registration process, thereby improving the accuracy of dose mapping accumulation in organ boundary areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 is a flow chart of an optional deformable registration method according to an embodiment of the present application;

[0024] Figure 2 is a schematic diagram of an optional data separation process according to an embodiment of the present application;

[0025] Figure 3 is a schematic diagram of an optional deformation registration process according to an embodiment of the present application;

[0026] Figure 4 is a schematic diagram of an optional deformation registration device according to an embodiment of the present application. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] According to an embodiment of the present application, a method embodiment of a deformable registration method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0030] In an optional embodiment, a deformable registration system can serve as the execution subject of the deformable registration method of the embodiment of the present application, wherein the deformable registration system can be a software system or an embedded system that combines software and hardware. In addition, in the embodiment of the present application, in addition to using the deformable registration system as the execution subject of the deformable registration method, other forms of execution subjects such as equipment and devices can also be used to execute the deformable registration method. It should be understood by those skilled in the art that the embodiment of the present application does not specifically limit the specific form of the method execution subject.

[0031] Figure 1 is a flow chart of an optional deformation registration method according to an embodiment of the present application, such as Figure 1 As shown, the method includes the following steps:

[0032] Step S101: Acquire reference data and floating data.

[0033] The reference data includes: a reference image, a delineated image of the reference image, and a dose map corresponding to the reference image; the floating data includes: a floating image, a delineated image of the floating image, and a dose map corresponding to the floating image.

[0034] Optionally, the reference data may be a set of image data used as a registration benchmark during the registration process, wherein the reference data includes: a reference image , the outline image of the reference image And the dose map corresponding to the reference image .

[0035] Among them, the outline image of the reference image Used to characterize the outline information of each organ in the reference image, and the dose map corresponding to the reference image Used to characterize the dose information corresponding to each organ area in the reference image.

[0036] Optionally, floating data refers to image data that needs to be registered with reference data, wherein the floating data includes: floating image , outline image of floating image Dose map corresponding to the floating image .

[0037] Among them, the outline image of the floating image Used to represent the outline information of each organ in the floating image, and the dose map corresponding to the floating image Used to characterize the dose information corresponding to each organ area in the floating image.

[0038] It should be noted that the outline image of the reference image and outline images of floating images The data stores the category and quantity information of the key tissue structures (independent organs) in radiotherapy. and sketch images There is a one-to-one correspondence between the number and category of outlines between the two.

[0039] Step S102 : performing data separation on the reference data and the floating data according to the organ type, wherein the data separation is used to obtain image data of each independent organ from the reference data and the floating data.

[0040] Optionally, first, the deformable registration system can be based on the outline image of the reference image The outline category of the tissue structure (independent organ) in the image is separated and extracted separately to obtain N independent outline reference data corresponding to different tissue structures (independent organs) (which can be recorded as ), where reference data are drawn independently Size and outline of the image Same, among which, .

[0041] Secondly, the deformable registration system can be based on the independent delineation reference data , the reference image Separate into N independent reference data Corresponding independent organ reference images .

[0042] In addition, after completing the image separation of tissue structures (independent organs), the deformation registration system can be used to Extract the reference image of the independent organ The image area outside the image is used as the background reference image , where the background reference image The way to obtain is .

[0043] From the above content, it can be seen that after separating the reference data according to organ type, independent delineation reference data can be obtained. , independent organ reference images and background reference image .

[0044] The data separation method is similar to that of the reference data. After the floating data are separated according to the organ type, the independent outline floating data can be obtained. , independent organ floating images and background floating images .

[0045] Step S103: performing deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ.

[0046] Optionally, the purpose of deformable registration is to align each independent organ in the floating image to the corresponding organ in the reference image so that the contours and internal structures of the organs match as closely as possible. During the deformable registration process, the image data of each independent organ will be deformed and registered separately, that is, the deformation of each organ can be calculated independently, avoiding mutual interference between different organs. In deformable registration, an optimization function can be used to measure the degree of match between the floating organ and the reference organ. By calculating the optimization function, a deformation field corresponding to each organ is generated. The deformation field is a vector field that can represent the direction and distance that each pixel point in the floating image moves during the deformation process. The deformation field can be used to guide the deformation of the floating image so that it is aligned with the corresponding organ in the reference image.

[0047] In step S104, the deformation fields corresponding to each independent organ are fused into a target deformation field, and deformation registration is performed on the floating data using the target deformation field.

[0048] Optionally, the target deformation field may be the complete deformation field of all tissue structures (independent organs), that is, the sum of the deformation fields of all tissue structures (independent organs).

[0049] In addition, the target deformation field can also be the total deformation field after adding the deformation field of the image background area and / or the deformation field of the overlapping area on the basis of the complete deformation field of the tissue structure (independent organ), where the overlapping area refers to the overlapping area between the image background area and the area where the organ is located.

[0050] Based on the above steps S101 to S104, we can see that this application splits the overall delineation and image information into multiple independent delineation and image information corresponding to different organs based on organ type, replacing the traditional global registration method. By performing independent registration for each organ, mutual interference between different organs is avoided, effectively solving the problem of redundant or insufficient alignment of local organ tissue structures during the registration process, thereby improving the accuracy of dose mapping accumulation in organ boundary areas.

[0051] In an optional embodiment, deformation registration is performed on the floating data through a target deformation field, including: determining a rigid transformation matrix based on a floating image; performing a rigid transformation on the floating data based on the rigid transformation matrix to obtain intermediate floating data, wherein the intermediate floating data includes a floating image after rigid transformation, an outline image, and a dose map; and deformation registration is performed on the intermediate floating data through a target deformation field.

[0052] Optionally, the deformable registration system can select a rigid registration model and then Perform rigid transformation and obtain rigid transformation matrix R in the process of rigid transformation. The deformation registration system can use rigid transformation matrix R to perform rigid transformation on floating data to obtain intermediate floating data. Among them, the intermediate floating data includes: floating image after rigid transformation , sketch the image and dose diagram .

[0053] In an optional embodiment, the reference data and floating data are separated based on organ type, including: the deformable registration system can separate each outlined area based on the outline information of each organ in the image, obtaining N independent outline reference data of the same size as the reference data and N independent outline floating data of the same size as the floating data. Then, the deformable registration system can segment the reference image into N independent organ reference images based on the N independent outline reference data corresponding to the reference data, and use the image area of ​​the reference image excluding the N independent organ reference images as the background reference image. Finally, the deformable registration system can segment the floating image into N independent organ floating images based on the N independent outline floating data corresponding to the floating data, and use the image area of ​​the floating image excluding the N independent organ floating images as the background floating image.

[0054] Optionally, when separating the reference data according to organ type, the following steps are included:

[0055] First, the deformable registration system can be used to define the outline image of the reference image. The outline category of the tissue structure (independent organ) in the image is separated and extracted separately to obtain N independent outline reference data corresponding to different tissue structures (independent organs) (which can be recorded as ), where reference data are drawn independently Size and outline of the image Same, among which, .

[0056] Secondly, the deformable registration system can be based on the independent delineation reference data , the reference image Separate into N independent reference data Corresponding independent organ reference images .

[0057] In addition, after completing the image separation of tissue structures (independent organs), the deformation registration system can be used to Extract the reference image of the independent organ The image area outside the image is used as the background reference image , where the background reference image The way to obtain is .

[0058] From the above content, it can be seen that after separating the reference data according to organ type, independent delineation reference data can be obtained. , independent organ reference images and background reference image .

[0059] Optionally, when separating the floating data according to organ type, the following steps are included:

[0060] First, the deformable registration system can be used to delineate the floating image. The outline category of the tissue structure (independent organ) in the image is separated and extracted separately, and N independent outline floating data corresponding to different tissue structures (independent organs) are generated (which can be recorded as ), where floating data are plotted independently Size and outline of the image Same, among which, .

[0061] Secondly, the deformable registration system can be based on the independent delineation of floating data , the floating image Separate into N independent floating data Corresponding independent organ floating images .

[0062] In addition, after completing the image separation of tissue structures (independent organs), the deformable registration system can be used to Extract floating images except for independent organs The image area outside the image is used as the background reference image , where the background reference image The way to obtain is .

[0063] From the above content, we can know that after separating the floating data according to organ type, we can get independent outline floating data , independent organ floating images and background floating images .

[0064] Optionally, Figure 2 is a schematic diagram of an optional data separation process according to an embodiment of the present application, such as Figure 2 As shown, based on the initial image outline information (such as the outline image of the reference image or the outline image of the floating image), through data separation, independent outline data 1, independent outline data 2, independent outline data 3, independent outline data 3, independent outline data 4 and independent outline data 6 can be separated, and each independent outline data corresponds to an independent organ.

[0065] In an optional embodiment, deformation registration is performed on the image data of each independent organ to obtain a deformation field corresponding to each independent organ, including: constructing a first function, a second function, and a third function, wherein the first function is used to correct the contour alignment error between the floating data and the reference data based on each organ; the second function is used to correct the contour registration redundancy between the floating data and the reference data based on each organ, and the contour registration redundancy represents that the contour of the registered organ exceeds the boundary range corresponding to the organ in the reference image; the third function is used to constrain the degree of deformation of each organ during the registration process. The deformation registration system can perform a weighted sum calculation on the first function, the second function, and the third function to obtain a target function, and then perform deformation registration on the image data of each independent organ according to the target function to obtain a deformation field corresponding to each independent organ.

[0066] Optionally, this application designs an optimization function for tissue structure (independent organ) deformation (i.e., the aforementioned objective function). This optimization function involves both the outline information corresponding to the image and the image information itself. The outline information corresponding to the image is used to ensure the precise alignment of the tissue structure (independent organ) contour, while the image information itself is used to constrain the rationality of the deformation and avoid folding. This application's optimization function is highly flexible and can be seamlessly embedded in other deformation registration optimization frameworks, demonstrating good reusability.

[0067] The optimization function of the present application includes: a first function, a second function, and a third function. Specifically, the optimization function can be a weighted sum of the first function, the second function, and the third function.

[0068] Optionally, the design goal of the first function is to correct image data with insufficient contour alignment during the registration process, the design goal of the second function is to correct image data with redundant contour alignment during the registration process, and the design goal of the third function is to constrain the degree of deformation of each organ during the registration process.

[0069] The phenomenon of "redundant or insufficient alignment" can be illustrated by the example of bladder registration: when the bladder meridian in the floating image is deformed and registered to the reference image, if the bladder contour exceeds the boundary of the bladder in the reference image after registration, it is "redundant alignment"; conversely, if it fails to completely cover the area of ​​the bladder in the reference image, it is "insufficient alignment".

[0070] In an optional embodiment, constructing a first function, a second function, and a third function includes: constructing a first function and a second function based on the independent outlined reference data and the independent outlined floating data corresponding to the i-th organ, wherein i is a positive integer, and the first function is used to determine the content overlap between the independent outlined reference data and the independent outlined floating data corresponding to the i-th organ; the second function is used to determine the contour redundancy of the independent outlined reference data corresponding to the i-th organ relative to the independent outlined floating data corresponding to the i-th organ, and determining the contour redundancy of the independent outlined floating data corresponding to the i-th organ relative to the independent outlined reference data corresponding to the i-th organ; constructing a third function based on the independent organ floating image and the independent organ reference image corresponding to the i-th organ, wherein the third function is used to determine the pixel overlap between the independent organ floating image and the independent organ reference image corresponding to the i-th organ.

[0071] Optionally, first, the first function is represented by the following formula (1).

[0072] (1)

[0073] In formula (1), the value calculated by the Dice function is used to measure the independent outline floating data and independently delineated reference data The degree of overlap between them.

[0074] In addition, it should be noted that in image data, a grid point is usually used to store a pixel value, which represents the grayscale or color information of the corresponding position on the image. In the deformation field, each grid point stores not a pixel value, but a two-dimensional or three-dimensional coordinate value, which is used to indicate the sampling position corresponding to the position in the data image. In other words, each grid point in the deformation field contains a mapping coordinate, which indicates the position in the data image from which the pixel value should be obtained. The “ "The symbol indicates that: according to the coordinate information in the deformation field, the corresponding position is found in the original image, and then the pixel value of the point is extracted to fill the current grid position. The " ” has the same meaning, so it will not be elaborated on later.

[0075] Next, the second function is represented by the following formula (2) and formula (3).

[0076] Among them, formula (2) is used to determine the contour redundancy of the independent outline floating data corresponding to the i-th organ relative to the independent outline reference data corresponding to the i-th organ; formula (3) is used to determine the contour redundancy of the independent outline reference data corresponding to the i-th organ relative to the independent outline floating data corresponding to the i-th organ. It corresponds to two cases of alignment redundancy: one is the redundancy of the floating tissue structure relative to the reference structure, and the other is the redundancy of the reference structure relative to the floating structure, respectively using parameters and Take measurements.

[0077] (2)

[0078] (3)

[0079] The Redun function is a redundancy metric used to measure the extent to which the outline of a floating structure overlaps with the reference structure outline beyond the necessary range during deformable registration. In other words, Redun helps identify and quantify instances where floating structures partially extend beyond or overlap the boundaries of the reference structure during registration. This is particularly important for evaluating the accuracy of registration algorithms and avoiding excessive deformation.

[0080] Furthermore, the deformation registration system can also use the mutual information index to measure the degree of pixel overlap between the floating tissue structure image and the reference tissue structure image to constrain the rationality of the deformation, as shown in formula (4):

[0081] (4)

[0082] Among them, formula (4) can represent the third function mentioned above. The MIM function usually refers to the "mutual information maximization" function. In the field of medical image processing and radiotherapy dose accumulation, it is an image registration method. Mutual information measures the dependence between two random variables and is used to evaluate the similarity between two images. It is particularly suitable for the registration of different types of images, such as CT images and MRI images, because the grayscale value ranges and physical meanings of these images may be different. In deformable registration, the mutual information function can help the algorithm understand and match the information between two images, even if there are significant differences in brightness or contrast. This is because mutual information focuses on the information sharing between the two images rather than directly comparing grayscale values. In dose accumulation calculations, the use of the mutual information function can ensure that the dose distribution can be accurately mapped to the reference image even when the anatomical structure changes, thereby improving the accuracy of dose assessment.

[0083] Then, through weighted summation, the optimization function (i.e., objective function) of the entire deformable registration is obtained, see formula (5).

[0084] (5)

[0085] In formula (5), 、 、 as well as They refer to 、 、 as well as The corresponding preset weight value.

[0086] Finally, according to the above optimization function, the intermediate floating data after rigid transformation (including the floating image after rigid transformation) can be sequentially , sketch the image and dose diagram ) to perform deformation registration and obtain the deformation field corresponding to each tissue structure (independent organ) ,in .

[0087] In an optional embodiment, the deformation field corresponding to each independent organ is fused into a target deformation field, including: adding and summing the deformation fields corresponding to each independent organ to obtain the organ deformation field; obtaining the deformation field of the image background area between the reference data and the floating data; obtaining the deformation field of the overlapping area between the image background area and the area where the organ is located; and summing the deformation field of the organ, the deformation field of the image background area, and the deformation field of the overlapping area to obtain the target deformation field.

[0088] Alternatively, first, because different tissue structures (independent organs) are highly independent and have almost no overlapping areas, all independent deformation fields can be combined. Direct fusion into a deformation field with complete tissue structure (i.e. the organ deformation field mentioned above), as shown in formula (6).

[0089] (6)

[0090] Secondly, the edge continuity between the image background area and the tissue structure (independent organ) is very strong, and there are a lot of overlapping areas. In order to ensure the accuracy of the tissue structure (independent organ) during registration, it is also necessary to obtain the deformation field of the image background area (which can be recorded as ) and the deformation field of the overlapping region (which can be expressed as ).

[0091] Finally, by combining the organ deformation field , the deformation field of the image background area and the deformation field in the overlapping region , the final target deformation field can be calculated based on the following formula (7): .

[0092] (7)

[0093] In an optional embodiment, obtaining a deformation field between the reference data and the floating data with respect to the image background region includes constructing a fourth function based on the background floating image and the background reference image, wherein the fourth function is used to determine a degree of deformation alignment between the background floating image and the background reference image. A deformation registration system performs deformation registration on the background floating image based on the fourth function to obtain a deformation field for the image background region.

[0094] Optionally, after the entire image is separated, the remaining image information is the background information image. The mutual information index can be directly used to measure the degree of deformation alignment of the image background area, as shown in formula (8):

[0095] (8)

[0096] According to the above formula (8), the deformation registration system can be used to transform the background floating image after rigid transformation. Perform deformation registration to obtain the deformation field of the image background area Wherein, formula (8) is used to represent the fourth function mentioned above.

[0097] In an optional embodiment, obtaining the deformation field of the overlapping region between the image background region and the region where the organ is located includes: a deformation registration system determines a mask corresponding to the overlapping region between the image background region and the region where the organ is located based on the intersection of the deformation field of the organ and the deformation field of the image background region; then extracts first deformation field data corresponding to the overlapping region from the deformation field of the organ based on the mask; and finally, extracts second deformation field data corresponding to the overlapping region from the deformation field of the image background region based on the mask. Finally, the deformation registration system performs a weighted sum calculation on the first deformation field data and the second deformation field data to obtain the deformation field of the overlapping region.

[0098] Optionally, as mentioned above, the edge continuity between the image background area and the tissue structure (independent organ) is very strong, and there are a large number of overlapping areas. In order to further improve the registration accuracy of the tissue structure (independent organ), this application mainly uses the image data of the tissue structure (independent organ) when performing data fusion of overlapping areas.

[0099] In addition, due to the complete deformation field of the tissue structure (independent organ) Deformation field of the image background area They are calculated independently, so the mask of the overlapping area can be obtained by finding the intersection of the two. Based on this mask, the values ​​of the corresponding areas can be extracted from the two deformation fields, and the deformation field fusion of the overlapping area can be completed accordingly. The specific process is shown in formula (9).

[0100] (9)

[0101] In an optional embodiment, Figure 3 is a schematic diagram of an optional deformation registration process according to an embodiment of the present application, such as Figure 3 As shown, the deformation registration process in the embodiment of the present application can be divided into four stages, namely: acquiring data, performing image decomposition based on tissue structure (independent organs), building a registration model (optimization function) based on the image decomposition data of the tissue structure, and performing deformation field fusion and deformation registration based on the image decomposition data of the tissue structure.

[0102] Alternatively, as Figure 3As shown, in the "data acquisition" stage, serial images of the patient during multiple treatments can be acquired, as well as outline images of core tissue structures (core independent organs) and serial treatment dose images generated based on these serial images.

[0103] Alternatively, as Figure 3 As shown in the figure, during the "Tissue Structure (Independent Organ) Image Decomposition" stage, the image to be decomposed and its outlines (reference data and floating data) are first input into the deformable registration system. The reference data includes the reference image, its outline image, and the dose map corresponding to the reference image; the floating data includes the floating image, its outline image, and the dose map corresponding to the floating image.

[0104] Then, the deformable registration system separates the reference data and floating data according to the organ type, and obtains N independent outline reference data of the same size as the reference data, and N independent outline floating data of the same size as the floating data; N independent organ reference images corresponding to the reference image, and N independent organ floating images corresponding to the floating image; 1 background reference image corresponding to the reference image, and 1 background floating image corresponding to the floating image.

[0105] It can be seen from this that after separating the floating data according to organ type, N independent outline floating data, N independent organ floating images and 1 background floating image can be obtained as decomposed floating information, and N independent outline reference data, N independent organ reference images and 1 background reference image can be obtained as decomposed reference information.

[0106] Alternatively, as Figure 3 As shown in FIG, in the stage of “building a registration model (optimization function) based on image decomposition data of tissue structure”, the deformable registration system can first select a rigid registration model to perform a rigid transformation on the floating data based on the decomposed reference information and the decomposed floating information, and then obtain the floating data after the rigid transformation (i.e., the intermediate floating data mentioned above), and then use the deformable registration model (optimization function) to determine the deformation fields corresponding to N tissue structures (independent organs) and the deformation field of the image background area.

[0107] Finally, if Figure 3 As shown in the figure, in the stage of "fusion and deformation registration of deformation fields based on image decomposition data of tissue structure", the deformation registration system uses the deformation field fusion strategy to fuse the deformation fields corresponding to N tissue structures (independent organs) and the deformation field of the image background area, obtains a complete deformation field and deforms the floating data, thereby obtaining the deformed floating image, independent organ outline image and dose map.

[0108] According to another aspect of the embodiment of the present application, a deformation registration device is further provided, wherein: Figure 4 is a schematic diagram of an optional deformation registration device according to an embodiment of the present application, such as Figure 4 As shown, the deformable registration device includes: an acquisition unit 401 , a data separation unit 402 , a first registration unit 403 , and a second registration unit 404 .

[0109] Among them, the acquisition unit 401 is used to acquire reference data and floating data, wherein the reference data includes: a reference image, a delineated image of the reference image, and a dose map corresponding to the reference image; the floating data includes: a floating image, a delineated image of the floating image, and a dose map corresponding to the floating image; the data separation unit 402 is used to perform data separation on the reference data and the floating data respectively according to the organ type, wherein the data separation is used to separate the image data of each independent organ from the reference data and the floating data; the first registration unit 403 is used to perform deformation registration on the image data of each independent organ to obtain the deformation field corresponding to each independent organ; the second registration unit 404 is used to fuse the deformation field corresponding to each independent organ into a target deformation field, and perform deformation registration on the floating data through the target deformation field.

[0110] Optionally, the second registration unit 404 includes: a first determination subunit, used to determine the rigid transformation matrix based on the floating image; a rigid transformation subunit, used to perform rigid transformation on the floating data based on the rigid transformation matrix to obtain intermediate floating data, wherein the intermediate floating data includes the floating image after the rigid transformation, the outline image and the dose map; and a deformation registration subunit, used to perform deformation registration on the intermediate floating data through the target deformation field.

[0111] Optionally, the data separation unit 402 includes: a data separation subunit, used to separate each outlined area separately according to the outline information of each organ in the image, and obtain N independent outlined reference data of the same size as the reference data and N independent outlined floating data of the same size as the floating data; an image segmentation subunit, used to segment the reference image into N independent organ reference images according to the N independent outlined reference data corresponding to the reference data, and use the image area of ​​the reference image other than the N independent organ reference images as the background reference image; and segment the floating image into N independent organ floating images according to the N independent outlined floating data corresponding to the floating data, and use the image area of ​​the floating image other than the N independent organ floating images as the background floating image.

[0112] Optionally, the first registration unit 403 includes: a function construction subunit, which is used to construct a first function, a second function and a third function, wherein the first function is used to correct the contour alignment error between the floating data and the reference data based on each organ; the second function is used to correct the contour registration redundancy between the floating data and the reference data based on each organ, and the contour registration redundancy characterizes that the contour of the organ after registration exceeds the boundary range corresponding to the organ in the reference image; the third function is used to constrain the degree of deformation of each organ during the registration process; a function calculation subunit, which is used to perform weighted sum calculation on the first function, the second function and the third function to obtain the target function; and the registration subunit, which is used to perform deformation registration on the image data of each independent organ according to the target function to obtain the deformation field corresponding to each independent organ.

[0113] Optionally, the function construction subunit includes: a first processing module, used to construct a first function and a second function based on the independent outline reference data and independent outline floating data corresponding to the i-th organ, wherein i is a positive integer, and the first function is used to determine the content overlap between the independent outline reference data and the independent outline floating data corresponding to the i-th organ; the second function is used to determine the contour redundancy of the independent outline reference data corresponding to the i-th organ relative to the independent outline floating data corresponding to the i-th organ, and determine the contour redundancy of the independent outline floating data corresponding to the i-th organ relative to the independent outline reference data corresponding to the i-th organ; the second processing module is used to construct a third function based on the independent organ floating image corresponding to the i-th organ and the independent organ reference image, wherein the third function is used to determine the pixel overlap between the independent organ floating image corresponding to the i-th organ and the independent organ reference image.

[0114] Optionally, the second registration unit 404 includes: a deformation field calculation subunit, which is used to add and sum the deformation fields corresponding to each independent organ to obtain the organ deformation field; a first acquisition subunit, which is used to obtain the deformation field of the image background area between the reference data and the floating data; a second acquisition subunit, which is used to obtain the deformation field of the overlapping area between the image background area and the area where the organ is located; and a deformation field summation calculation subunit, which is used to sum and calculate the deformation field of the organ, the deformation field of the image background area, and the deformation field of the overlapping area to obtain the target deformation field.

[0115] Optionally, the first acquisition subunit includes: a third processing module, used to construct a fourth function based on the background floating image and the background reference image, wherein the fourth function is used to determine the deformation alignment between the background floating image and the background reference image; a fourth processing module, used to perform deformation registration on the background floating image according to the fourth function to obtain the deformation field of the image background area.

[0116] Optionally, the second acquisition subunit includes: a fifth processing module, used to determine the mask corresponding to the overlapping area between the image background area and the area where the organ is located based on the intersection between the organ deformation field and the deformation field of the image background area; a sixth processing module, used to extract first deformation field data corresponding to the overlapping area from the organ deformation field according to the mask; a seventh processing module, used to extract second deformation field data corresponding to the overlapping area from the deformation field of the image background area according to the mask; and an eighth processing module, used to perform weighted sum calculation on the first deformation field data and the second deformation field data to obtain the deformation field of the overlapping area.

[0117] According to another aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored. When the computer program runs, the device where the computer-readable storage medium is located executes the above-mentioned deformation registration method.

[0118] According to another aspect of an embodiment of the present application, an electronic device is also provided, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors execute the above-mentioned deformation registration method.

[0119] The serial numbers of the above embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0120] In the above embodiments of the present application, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0121] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0122] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0123] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0124] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk, etc. Various media that can store program code.

[0125] The above is only a preferred embodiment of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A deformation registration method, characterized in that: include: Acquiring reference data and floating data, wherein the reference data includes: a reference image, a delineated image of the reference image, and a dose map corresponding to the reference image; the floating data includes: a floating image, a delineated image of the floating image, and a dose map corresponding to the floating image; performing data separation on the reference data and the floating data respectively according to organ type, wherein the data separation is used to separate the image data of each independent organ from the reference data and the floating data; Performing deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ; The deformation fields corresponding to each independent organ are fused into a target deformation field, and deformation registration is performed on the floating data using the target deformation field.

2. The deformable registration method according to claim 1, characterized in that: Performing deformation registration on the floating data using the target deformation field includes: determining a rigid transformation matrix based on the floating image; Performing a rigid transformation on the floating data according to the rigid transformation matrix to obtain intermediate floating data, wherein the intermediate floating data includes a floating image, a delineated image, and a dose map after the rigid transformation; Deformation registration is performed on the intermediate floating data using the target deformation field.

3. The deformable registration method according to claim 1, wherein: Separating the reference data and the floating data according to organ type includes: Separating each outlined area according to the outline information of each organ in the image to obtain N independent outlined reference data of the same size as the reference data and N independent outlined floating data of the same size as the floating data; Segmenting the reference image into N independent organ reference images according to N independent delineation reference data corresponding to the reference data, and using an image region of the reference image other than the N independent organ reference images as a background reference image; The floating image is divided into N independent organ floating images according to the N independent outline floating data corresponding to the floating data, and the image area of ​​the floating image except the N independent organ floating images is used as a background floating image.

4. The deformable registration method according to claim 1, wherein: Performing deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ includes: Constructing a first function, a second function, and a third function, wherein the first function is used to correct the contour alignment error between the floating data and the reference data based on each organ; the second function is used to correct the contour registration redundancy between the floating data and the reference data based on each organ, the contour registration redundancy representing that the contour of the registered organ exceeds the boundary range corresponding to the organ in the reference image; and the third function is used to constrain the degree of deformation of each organ during the registration process; Performing weighted sum calculation on the first function, the second function, and the third function to obtain an objective function; Deformation registration is performed on the image data of each independent organ according to the objective function to obtain a deformation field corresponding to each independent organ.

5. The deformable registration method according to claim 4, characterized in that: Construct the first, second, and third functions, including: Constructing the first function and the second function according to the independent delineation reference data and the independent delineation floating data corresponding to the i-th organ, wherein i is a positive integer, and the first function is used to determine the content overlap between the independent delineation reference data and the independent delineation floating data corresponding to the i-th organ; and the second function is used to determine the contour redundancy of the independent delineation reference data corresponding to the i-th organ relative to the independent delineation floating data corresponding to the i-th organ, and to determine the contour redundancy of the independent delineation floating data corresponding to the i-th organ relative to the independent delineation reference data corresponding to the i-th organ; The third function is constructed based on the independent organ floating image and the independent organ reference image corresponding to the i-th organ, wherein the third function is used to determine the pixel coincidence between the independent organ floating image and the independent organ reference image corresponding to the i-th organ.

6. The deformable registration method according to claim 1, wherein: The deformation fields corresponding to each independent organ are fused into a target deformation field, including: Adding and summing the deformation fields corresponding to each independent organ to obtain the organ deformation field; Acquire a deformation field between the reference data and the floating data with respect to an image background area; Acquire a deformation field of an overlapping area between the image background area and the area where the organ is located; The target deformation field is obtained by summing up the deformation field of the organ, the deformation field of the image background area, and the deformation field of the overlapping area.

7. The deformable registration method according to claim 6, characterized in that: Acquiring a deformation field between the reference data and the floating data with respect to an image background area, comprising: constructing a fourth function according to the background floating image and the background reference image, wherein the fourth function is used to determine a deformation alignment between the background floating image and the background reference image; Deformation registration is performed on the background floating image according to the fourth function to obtain a deformation field of the image background area.

8. The deformable registration method according to claim 6, characterized in that: Obtaining a deformation field of an overlapping area between the image background area and the area where the organ is located, including: Determining a mask corresponding to an overlapping area between the image background area and the area where the organ is located based on an intersection between the deformation field of the organ and the deformation field of the image background area; extracting first deformation field data corresponding to the overlapping area from the organ deformation field according to the mask; extracting second deformation field data corresponding to the overlapping area from the deformation field of the image background area according to the mask; The first deformation field data and the second deformation field data are weightedly summed to obtain a deformation field of the overlapping area.

9. A deformation registration device, characterized in that: include: an acquisition unit, configured to acquire reference data and floating data, wherein the reference data includes: a reference image, a delineated image of the reference image, and a dose map corresponding to the reference image; and the floating data includes: a floating image, a delineated image of the floating image, and a dose map corresponding to the floating image; a data separation unit, configured to perform data separation on the reference data and the floating data according to organ type, wherein the data separation is configured to separate image data of each independent organ from the reference data and the floating data; a first registration unit, configured to perform deformation registration on the image data of each independent organ to obtain a deformation field corresponding to each independent organ; The second registration unit is configured to fuse the deformation fields corresponding to each independent organ into a target deformation field, and perform deformation registration on the floating data using the target deformation field.

10. A computer-readable storage medium, characterized in that in, A computer program is stored in the computer-readable storage medium, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the deformable registration method according to any one of claims 1 to 8.

11. An electronic device, characterized in that: The method comprises one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the deformable registration method described in any one of claims 1 to 8.

Images