Providing posture information on X-ray projection images

Abstract

A computer-implemented method for providing pose information for an x-ray projection image is provided, the method comprising: 1..m segmenting the X-ray projection image 120 to identify the plurality of projected sub-regions 150 within the segmented X-ray projection image 120; 1..m generating a posture metric for the X-ray projection image 120 based on the relative sizes of two or more projected sub-regions of the X-ray projection image 120; and outputting the posture metric to provide posture information for the X-ray projection image 120.

Description

[Technical Field] 【0001】 SUMMARY A computer-implemented method, computer program product, and system are disclosed that provide posture information for x-ray projection images. [Background technology] 【0002】 A projection X-ray imaging system includes an X-ray source and an X-ray detector. The X-ray source and the X-ray detector are separated by an examination region. An anatomical structure, such as an ankle, leg, or another part of a subject's body, can be placed in the examination region to generate an X-ray projection image of the anatomical structure. The X-ray projection image is acquired by the projection X-ray imaging system having a defined orientation relative to the anatomical structure. The X-ray projection image is acquired using a single orientation of the X-ray imaging system relative to the anatomical structure, and therefore, it is important that the orientation used provides the desired information in the resulting X-ray projection image. For example, diagnosing a fractured ankle bone may require imaging the bone using an orientation that does not obscure the fracture by other bones in the ankle. If an improper orientation is used, repeated images may need to be acquired. Summary of the Invention [Problem to be solved by the invention] 【0003】 Positioning of an anatomical structure relative to a projection X-ray imaging system is traditionally performed manually and according to a protocol for the anatomical structure. The radiographer visually positions the anatomical structure and generates an initial image. If the initial image is unacceptable, the radiographer repositions the anatomical structure based on their experience and generates another image. This procedure can be repeated several times until an acceptable image is obtained, but sometimes the resulting image only improves slightly. Therefore, this approach increases the amount of X-ray dose delivered to the subject and also hinders workflow. 【0004】 US Patent Application Publication No. 2019 / 183438 A1 describes a method for ensuring accurate positioning for radiography recording, the method including providing an examination request for a body region, pre-positioning the body region in a radiography system for radiography recording, pre-positioning at least one of a recording unit of the radiography system and an image detector of the radiography system for radiography recording, generating a positioning record of the body region via the radiography system, where the radiography system is switched to a fluoroscopy mode and the positioning record is a fluoroscopy record, generating position information from the positioning record, and outputting the position information. 【0005】 Another document, WO 2020 / 038917 A1, relates to determining imaging orientation based on 2D projection images. Based on deep neural networks and possibly active shape modeling approaches, the complete contours of anatomical structures can be determined and classified. Based on specific features, the algorithm can evaluate whether the C-arm's viewing angle is sufficient. In another step, the algorithm can estimate how far the current viewing angle is from the desired viewing angle and provide guidance on how to adjust the C-arm's position to reach the desired viewing angle. 【0006】 However, there remains a need for improved positioning of anatomical structures relative to projection x-ray imaging systems. [Means for solving the problem] 【0007】 According to one aspect of the present disclosure, there is provided a computer-implemented method for providing pose information for an X-ray projection image, the method including: receiving X-ray projection data, the X-ray projection data including X-ray projection images representing an anatomical structure, the X-ray projection data being acquired by a projection X-ray imaging system having a corresponding pose relative to the anatomical structure; segmenting the X-ray projection image to identify a plurality of projected sub-regions of the anatomical structure; generating a pose metric for the X-ray projection image based on relative sizes of two or more of the projected sub-regions in the segmented X-ray projection image; and outputting the pose metric to provide the pose information for the X-ray projection image. 【0008】 In the above method, a posture metric is generated for an X-ray projection image. The posture metric is generated based on the relative sizes of two or more projected subregions within a segmented X-ray projection image. The inventors have observed that this relative size serves as a reliable metric of the posture of a projection X-ray imaging system relative to an anatomical structure. An operator can use the posture metric for various purposes, such as to determine the posture of the projection X-ray imaging system relative to an anatomical structure, to evaluate the suitability of the posture for acquiring X-ray projection images, or to determine how to adjust the posture of the projection X-ray imaging system to acquire improved X-ray projection images of the anatomical structure. Thus, the method facilitates reducing the X-ray dose delivered to a subject by reducing the number of repeated X-ray image acquisitions. The method also facilitates improved workflow. 【0009】 Other aspects, features, and advantages of the present disclosure will become apparent from the following description of embodiments that refers to the accompanying drawings. [Brief explanation of the drawings] 【0010】 [Figure 1]1 illustrates an example of an X-ray projection image 120 depicting an anatomical structure, according to some aspects of the present disclosure. [Figure 2] 1 is a flowchart illustrating an example of a method for providing pose information regarding an X-ray projection image, according to some aspects of the present disclosure. [Figure 3] 2 is a schematic diagram illustrating an example of a system 200 for providing pose information related to an X-ray projection image, in accordance with some aspects of the present disclosure. [Figure 4] 1 is a schematic diagram illustrating an example of a projection X-ray imaging system 140 having a pose P relative to an anatomical structure 130, in accordance with some aspects of the present disclosure. [Figure 5] FIG. 1 illustrates an example of a segmented x-ray projection image 120, including example projected sub-regions 1501..m of an anatomical structure, according to some embodiments of the present invention. [Figure 6] 6A-6C show three examples of segmented X-ray projection images 1201..3 according to some embodiments of the present invention, which include examples of projected sub-regions 1501..m of an anatomical structure and which were acquired from different poses P1 (FIG. 6a), P2 (FIG. 6b), and P3 (FIG. 6c) of a projection X-ray imaging system relative to the anatomical structure. [Figure 7] FIG. 1 illustrates an example of a segmented X-ray projection image 120, including example projected sub-regions 1501..m of an anatomical structure and example locations of multiple anatomical landmarks 1601..n, according to some aspects of the present disclosure. [Figure 8] FIG. 1 is a schematic diagram illustrating an example of a method for providing posture information for an X-ray projection image, in which X-ray projection data 110 including an X-ray projection image is input to a first neural network NN1 to segment the X-ray projection image, and segmented image data representing the segmented image is input to a second neural network NN2 to generate a posture metric for the X-ray projection image, in accordance with some aspects of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION 【0011】 Examples of the present disclosure are provided with reference to the following description and drawings. In this detailed description, for purposes of explanation, many specific details of particular examples are set forth. Reference herein to an "example," "embodiment," or similar language means that a feature, structure, or characteristic described in connection with an example is included in at least that example. Furthermore, it should be understood that features described in connection with one example may also be used in other examples, and that for purposes of brevity, not all features are necessarily repeated in each example. For example, features described in connection with a computer-implemented method may be implemented in a corresponding manner in a system and in a computer program product. 【0012】 In the following description, reference will be made to examples of how a projection X-ray imaging system is used to obtain X-ray projection images representing anatomical structures. As some examples, the projection X-ray imaging system can be a DigitalDiagnost C90, or a Philips Azurion 7, or a MobileDiagnost M50 mobile X-ray system, all of which are commercially available from Philips Healthcare (Best, the Netherlands). However, it should be understood that these serve as examples only, and that the projection X-ray imaging system can generally be provided by any type of projection X-ray imaging system. 【0013】 In some examples, configurations of X-ray imaging systems are described in which the X-ray source of a projection X-ray imaging system is mounted to the ceiling via a gantry and the corresponding X-ray detector is mounted to a stand and held in a vertical position. This type of configuration can be used with the DigitalDiagnost C90 imaging system described above. However, it should be understood that this configuration is merely an example, and that the X-ray source and X-ray detector of a projection X-ray imaging system can alternatively be configured differently. For example, the X-ray source can be mounted to a stand or an articulating arm, and the X-ray detector can be mounted to the ceiling or an articulating arm and held in a different position. Alternatively, configurations can be used in which the X-ray source and X-ray detector are mounted to a common support structure. Such a configuration is used in the Philips Azurion 7 projection X-ray imaging system described above. In this example, the support structure is a so-called C-arm. Alternatively, other configurations can be used in which the X-ray source and X-ray detector are mounted to a common support structure having a shape different from a C-arm, including configurations in which the support structure is provided by a so-called "O-arm." Alternatively, other configurations can be used in which the X-ray source and X-ray detector are mounted or supported differently, including, for example, portable X-ray imaging systems such as the MobileDiagnost M50 mobile digital X-ray system mentioned above. 【0014】 In the following description, reference is made to an example method involving the acquisition of X-ray projection images representing an anatomical structure. In some examples, the anatomical structure is an ankle. However, it should be understood that the ankle serves merely as an example, and that in general, the method disclosed herein can be used to acquire X-ray projection images representing any anatomical structure within a subject's body. While the following description refers to an example method in which a posture metric is generated for a single X-ray projection image, it should also be understood that the method can similarly be used to generate a posture metric for each of multiple images. The method can be used, for example, to provide a posture metric for a time sequence of images, where the posture metric changes in response to changes in posture of the projection X-ray imaging system during the sequence. 【0015】 In the following description, reference is made to various computer-implemented, i.e., processor-implemented, methods. It should be noted that the computer-implemented methods disclosed herein may be provided as a non-transitory computer-readable storage medium storing computer-readable instructions that, when executed by at least one processor, cause the at least one processor to perform the method. In other words, the computer-implemented methods may be embodied in a computer program. The computer program may be provided by dedicated hardware or hardware capable of executing software in association with appropriate software. When provided by a processor, the functionality of the method features may be provided by a single dedicated processor, by a single shared processor, or by multiple individual processors, some of which may be shared. Explicit use of the terms "processor" or "controller" should not be construed as exclusively referring to hardware capable of executing software, but may implicitly include, but is not limited to, digital signal processor (DSP) hardware, read-only memory (ROM) for storing software, random access memory (RAM), non-volatile storage, etc. Furthermore, examples of the present disclosure may take the form of a computer program product accessible from a computer-usable or computer-readable storage medium, the computer program product providing program code for use by or in connection with a computer or any instruction execution system. For purposes of this description, a computer-usable or computer-readable storage medium can be any device that can have, store, communicate, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system or device or propagation medium.Examples of computer-readable media include semiconductor or solid-state memory, magnetic tape, removable computer diskettes, random access memory "RAM," read-only memory "ROM," rigid magnetic disks, and optical disks, current examples of which include compact disk-read-only memory "CD-ROM," compact disk-read / write "CD-R / W," Blu-ray, and DVD. 【0016】 As noted above, there remains a need for improvements in the positioning of anatomical structures relative to projection x-ray imaging systems. 【0017】 FIG. 1 is an example of an X-ray projection image 120 depicting an anatomical structure, according to some embodiments of the present disclosure. A radiographer may acquire an image such as the image shown in FIG. 1 during a clinical study, for example, for the presence of a fractured bone in the ankle. However, as discussed above, a drawback of such projection X-ray images is that they must be acquired using a projection X-ray imaging system with an appropriate pose relative to the anatomical structure. This is a result of projection X-ray images being generated by projecting attenuation resulting from a three-dimensional object onto the surface of a two-dimensional X-ray detector. If the projection X-ray imaging system used to acquire the image shown in FIG. 1 has an improper pose with respect to the ankle, the fracture may be obscured by other bones in the ankle. This creates the need to acquire another image, thereby increasing the X-ray dose delivered to the subject and disrupting workflow. 【0018】 2 is a flowchart illustrating an example of a method for providing posture information for an X-ray projection image according to some aspects of the present disclosure. FIG. 3 is a schematic diagram illustrating an example of a system 200 for providing posture information for an X-ray projection image according to some aspects of the present disclosure. The system 200 includes one or more processors 210. The processing described in connection with the method illustrated in FIG. 2 may also be performed by the one or more processors 210 of the system 200 illustrated in FIG. 3. Similarly, the processing described in connection with the one or more processors 210 of the system 200 may also be performed in the method described with reference to FIG. 2. Referring to FIG. 2, a computer-implemented method for providing posture information for an X-ray projection image includes receiving (S110) X-ray projection data 110, the X-ray projection data including an X-ray projection image 120 representing an anatomical structure 130, the X-ray projection data having been acquired by a projection X-ray imaging system 140 having a corresponding posture P relative to the anatomical structure 130; and segmenting the X-ray projection image 120 to identify a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m and identifying (S120) the projected sub-regions 150 in the segmented X-ray projection image 120. 1..m generating a posture metric for the X-ray projection image 120 based on the relative sizes of two or more of the projected sub-regions (S130); and outputting the posture metric to provide posture information for the X-ray projection image 120 (S140). 【0019】 In the above method, a posture metric is generated for an X-ray projection image. The posture metric is generated based on the relative sizes of two or more projected subregions within a segmented X-ray projection image. The inventors have observed that this relative size serves as a reliable metric of the posture of a projection X-ray imaging system relative to an anatomical structure. An operator can use the posture metric for various purposes, such as to determine the posture of the projection X-ray imaging system relative to an anatomical structure, to evaluate the suitability of the posture for acquiring X-ray projection images, or to determine how to adjust the posture of the projection X-ray imaging system to acquire improved X-ray projection images of the anatomical structure. Thus, the method facilitates reducing the X-ray dose delivered to a subject by reducing the number of repeated X-ray image acquisitions. The method further facilitates improved workflow. 【0020】 2, in operation S110, x-ray projection data 110 is received. The x-ray projection data includes x-ray projection images 120 representing an anatomical structure 130. The x-ray projection data is acquired by a projection x-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130. 【0021】 A projection X-ray imaging system includes an X-ray source and an X-ray detector. The X-ray source and the X-ray detector are held in a stationary position relative to the anatomical structure during acquisition of the X-ray projection data. The projection X-ray imaging system can be said to have a posture relative to the anatomical structure during acquisition of the X-ray projection data. The X-ray projection data is acquired using a single posture of the projection X-ray imaging system relative to the anatomical structure. The X-ray projection data can be used to generate X-ray projection images, i.e., two-dimensional images, representing the anatomical structure. The X-ray projection data acquired by a projection X-ray imaging system is in contrast to volumetric X-ray data generated by a computed tomography (CT) imaging system. In a CT imaging system, the X-ray source and the X-ray detector rotate around the anatomical structure to acquire volumetric X-ray data from each of multiple postures of the CT imaging system relative to the anatomical structure. The volumetric X-ray data is then reconstructed into a three-dimensional, or volumetric, image of the anatomical structure. 【0022】 The X-ray projection data 110 received in process S110 can be received from a variety of sources. For example, the X-ray projection data 110 can be received from a projection X-ray imaging system, such as the projection X-ray imaging system 140 shown in FIG. 3. In this example, the X-ray projection data 110 can be received by one or more processors 210 of the system 200 shown in FIG. 3. Instead of being received from a projection X-ray imaging system, the X-ray projection data 110 can be received from another source, such as a computer-readable storage medium, the Internet, or the cloud. Thus, the X-ray projection data 110 can be acquired by the projection X-ray imaging system at an earlier point in time and then received from the computer-readable storage medium, the Internet, or the cloud in process S110. In general, the X-ray projection data 110 received in process S110 can be received via any form of data communication, including wired, optical, and wireless communication. As some examples, when wired or optical communication is used, communication can occur via signals transmitted over electrical or optical cables, and when wireless communication is used, communication can occur via RF or optical signals. 【0023】 As described above, the X-ray projection data 110 is acquired by a projection X-ray imaging system 140 having a corresponding orientation P with respect to the anatomical structure 130. FIG. 4 is a schematic diagram illustrating an example of a projection X-ray imaging system 140 having an orientation P with respect to the anatomical structure 130, according to some embodiments of the present disclosure. The projection X-ray imaging system 140 illustrated in FIG. 4 corresponds to the projection X-ray imaging system 140 illustrated in FIG. 3 and includes an X-ray source 140S and an X-ray detector 140D. An example of an anatomical structure is further illustrated in FIGS. 3 and 4 as an ankle. In both FIGS. 3 and 4, X-ray projection information of the ankle is acquired by detecting X-rays emitted from the X-ray source 140S and passing through the ankle using the X-ray detector 140D. The X-ray projection data represents a line integral of the X-ray attenuation of the ankle along a path between the X-ray source 140S and the X-ray detector 140D. 【0024】 The orientation P of the X-ray imaging system 140 relative to the ankle is indicated through the orientation of the thick, dark arrow in FIG. 4 relative to the ankle. In the illustrated example, the orientation P is defined by the orientation of the centerline of the projection X-ray imaging system 140 relative to the anatomical structure. The diagram within the dashed box in the upper right corner of FIG. 4 shows a generalized orientation P of the projection X-ray imaging system 140 relative to the anatomical structure 130. In this inset, the orientation P is defined by angles α and β relative to the coordinate system of the anatomical structure 130. As an example, angle α can be defined as the rotation angle of the centerline of the projection X-ray imaging system 140 about the longitudinal axis of the object, and angle β can be defined as the tilt angle of the centerline of the projection X-ray imaging system 140 relative to the craniocaudal axis of the object. 【0025】 Returning to the method shown in FIG. 2, in step S120, the x-ray projection image 120 is divided into a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m FIG. 5 is an example of a segmented x-ray projection image 120 showing projected sub-regions 150 of an anatomical structure, according to some embodiments of the present invention. 1..m In the illustrated example, the segmentation results in a projected sub-region of the ankle 150 1..m These include a first projected area of ​​the tibia 1501, a second projected area of ​​the tibia 1502, a projected joint space 1503 between the tibia and talus, a first articular surface 1504 of the talus, an overlap 1505 between the fibula and talus, and a second articular surface 1506 of the talus. 【0026】 Generally, the projected sub-region 150 of the anatomical structure 130 identified by the segmentation process 120 is 1..m Part of the bone 150 1..2 , 150 4..6, and the space 1503 between the two bones. Generally, a portion of a bone can include a diaphysis, i.e., shaft portion, or a metaphysis, i.e., epiphysis portion, of a bone. Examples of portions of a bone that can be identified as a projected sub-region in process S120 include protrusions on the bone's periphery and protrusions on the articular surface of a bone, such as protrusions on a bone facet (articular cavity). In the embodiment shown in FIG. 5, the projected sub-region 150 1..m is the tibia part, i.e. 150 1..2 , a portion of the fibula, i.e., 1505, and a portion of the talus, i.e., 1504 and 1506. Items 1504 and 1506 in FIG. 5 are projections of bone articular surfaces, i.e., a first articular surface of the talus 1504 and a second articular surface of the talus 1506. As mentioned above, the projected sub-regions can alternatively represent the space between two bones 1503. The space between two bones is also referred to herein as a gap, or more specifically, a joint gap. The joint gap is the gap associated with the articulation between two adjacent bones. The joint can be, for example, a synovial joint. In this case, at least a portion of the joint gap represented in the projected sub-region represents a portion of the synovial cavity. The portion of the joint gap represented in the projected sub-region can also represent cartilage loss due to degenerative joint disease or wear and tear. The joint space has a relatively lower attenuation of x-rays than bone, resulting in a region in the x-ray projection image that is relatively lower intensity than the bones on either side of the space. Item 1503 in Figure 5 is an example of a projection of the space between two bones, i.e., the projected joint space 1503 between the tibia and talus. 【0027】 Projected sub-area 150 1..m can be defined in various ways. In one embodiment, at least one of the two or more projected sub-regions is located in the x-ray projection image 150 1..6In another example, at least one of the two or more projected subregions is defined at least in part by the perimeter of a portion of at least one bone within the x-ray projection image. For example, in FIG. 5, projected subregion 1501 is defined, at least at its upper, left, and lower ends, by the perimeter of a bone facet, more specifically, the articular surface of the tibia. Similarly, projected subregion 1502 is defined, at least at its upper, left, and lower ends, by the perimeter of a bone facet, more specifically, the articular surface of the tibia. In another example, at least one of the two or more projected subregions is defined at least in part by the intersection of multiple bones (i.e., bone-to-bone) within the x-ray projection image. For example, in FIG. 5, projected subregion 1501 is defined, at least at its right end, by the intersection of the fibula and tibia. Similarly, projected subregion 1502 is defined, at least at its left end, by the intersection of the fibula and tibia. In another embodiment, at least one of the two or more projected sub-regions is defined at least in part by an intersection of the space 1503 between the two bones with another bone in the x-ray projection image 120. Although this is not explicitly shown in Figure 5, the projected sub-region according to this embodiment can be defined by an intersection of the joint space 1503 with the fibula. The projected sub-region can also be defined by a combination of these examples. 【0028】 A plurality of projected subregions 150 of the anatomical structure 130 1..mProcess S120, in which the X-ray projection images 120 are segmented to identify the anatomical structures 130, can be performed using a variety of techniques. One exemplary technique involves applying a segmentation algorithm to the received X-ray projection data 110. Segmentation algorithms such as model-based segmentation, watershed-based segmentation, region growing, level setting, or graph cut can be used for this purpose. Another exemplary technique involves inputting the received X-ray projection data 110 into a neural network trained to segment X-ray projection images representing the anatomical structures 130. An example of such a neural network is the first neural network NN1 described below with reference to FIG. 8. An example of a neural network that can be used to segment X-ray projection images in process S120 is Kroenke, S., et al., "CNN-based pose estimation for assessing quality of ankle-joint X-ray images," Proc. SPIE 12032, Medical Imaging 2022: Image Processing, 120321A (2022). 【0029】 Referring back to the method shown in FIG. 2, in step S130, the pose metric of the X-ray projection image 120 is calculated based on the projected sub-regions 150 in the segmented X-ray projection image 120. 1..m The segmented x-ray projection image 120 is generated based on the relative size of two or more projected sub-regions of the projection x-ray image 120. As mentioned above, the inventors have observed that this relative size serves as a reliable metric of the pose of the projection x-ray imaging system relative to the anatomical structure. This principle is used in the segmented x-ray projection image 120 according to some aspects of the present disclosure. 1..3 Three examples of segmented X-ray projection images are illustrated in FIG. 6, where the segmented X-ray projection images are divided into projected sub-regions 150 of the anatomical structure. 1..m6a), P2 (FIG. 6b), and P3 (FIG. 6c) are taken from different positions of the projection X-ray imaging system relative to the anatomy. 【0030】 At the top of FIG. 6, the dashed box indicates by a thick dark arrow the orientation P of the projection X-ray imaging system relative to the anatomical structure 130. 1..3 Attitude P 1..3 is the corresponding X-ray projection image 120 shown at the bottom of FIG. 1..3 The projection X-ray imaging system 140 is used to obtain the orientation P 1..3 is defined as described above with reference to FIG. 1..3 is defined by the orientation of the centerline of the projection X-ray imaging system relative to the anatomical structure. 1..3 can be defined by angles α and β relative to the coordinate system of the anatomy 130. Angle α can be defined as the angle of rotation of the centerline of the projection X-ray imaging system about the longitudinal axis of the object, and angle β can be defined as the angle of tilt of the centerline of the projection X-ray imaging system relative to the craniocaudal axis of the object. 【0031】 Referring first to FIG. 6b, in this position, pose P2 is optimal for generating an X-ray projection image representing the ankle. In this optimal pose, the tip of the fibula is centrally aligned with the tibia, and the projected joint space 1503 between the tibia and talus appears continuous along the length of the tibia's articular surface. In FIG. 6b, the following projected sub-regions are visible in optimal pose P2: a first projected region 1501 of the tibia, a second projected region 1502 of the tibia, the projected joint space 1503 between the tibia and talus, and the overlap 1505 between the fibula and talus. For each of these projected sub-regions, a size can be calculated. The size can be an area or distance measurement for each projected sub-region. As some examples, the area can be calculated as the total area of ​​the projected sub-regions or as the area of ​​a shape fitted to the projected sub-regions, the distance measure can be calculated as the maximum, minimum, or average distance of the projected sub-regions or the distance measure of a shape fitted to the projected sub-regions, the size can be calculated in units of (square) millimeters, (square) centimeters, etc., or in another reference unit such as image pixels. 【0032】 Referring now to FIG. 6a, in this illustration, pose P1 is obtained from pose P2 by rotating the projection X-ray imaging system counterclockwise about 5 degrees around the y-axis of the anatomy. Pose P1 is not optimal for generating an X-ray projection image representing the ankle. This is due in part to the fact that the projected joint space 1503 between the tibia and talus is no longer continuous along the length of the tibial articular surface. Compared to FIG. 6b, in FIG. 6a, the size of the first projected region 1501 of the tibia has increased, while the size of the second projected region 1502 of the tibia has decreased. The size of the overlap between the fibula and talus 1505 has also decreased. Thus, rotating the projection X-ray imaging system counterclockwise about the y-axis of the anatomy to obtain pose P2 has resulted in changes in the sizes of the projected subregions 1501, 1502, 1503, and 1505. 【0033】 Referring now to Figure 6c, posture P3 was obtained from posture P2 by rotating the projection X-ray imaging system approximately 5 degrees clockwise around the y-axis of the anatomy. Pose P3 is not optimal for generating an X-ray projection image of the ankle. This is due in part to the fact that the projected joint space 1503 between the tibia and talus is no longer continuous along the length of the tibial articular surface. Compared to Figure 6b, in Figure 6a, the size of the first projection region 1501 of the tibia has decreased, while the size of the second projection region 1502 of the tibia has increased. The size of the overlap 1505 between the fibula and talus has also decreased. Thus, rotating the projection X-ray imaging system clockwise around the y-axis of the anatomy to obtain posture P3 has also resulted in changes in the sizes of the projected subregions 1501, 1502, 1503, and 1505. 【0034】 As can be seen from the above description, the projected sub-region 150 with rotation about the y-axis 1..m As a result of the change in size of the projected sub-region 150 in the segmented X-ray projection image 120, 1..m The relative sizes of two or more of the projected sub-regions of 1..3 For example, the relative size of projected sub-region 1501 compared to the size of projected sub-region 1502 provides a measure of the rotation of the ankle about the y-axis of the anatomy. The relative size of projected sub-region 1504 compared to the size of projected sub-region 1503 also provides a measure of the rotation of the ankle about the y-axis of the anatomy. Other orientations of the projection X-ray imaging system relative to the anatomy 130 can be determined in a similar manner by adjusting the size of projected sub-region 1504. 1..m can be determined from the relative sizes of 【0035】 2 or more projected subregions 150 to generate pose metrics 1..mVarious examples of process S130 in which the relative sizes of the projected sub-regions are used are described below, including one approach in which the sizes of the projected sub-regions are calculated from the X-ray projection images, and another approach in which the X-ray projection images 120 are input to a neural network, which is trained to generate a pose metric using ground truth values ​​of the pose metric calculated based on the relative sizes of two or more of the projected sub-regions in the segmented X-ray projection images. 【0036】 In some examples, the posture metric represents the posture P of the projection X-ray imaging system 140 relative to the anatomical structure 130. In some examples, the posture metric represents the suitability of the posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120. In some examples, the posture metric represents feedback for adjusting the posture P of the projection X-ray imaging system for acquiring a subsequent X-ray projection image representing the anatomical structure 130. The posture metric may also represent a combination of these examples. The subsequent X-ray projection image may be referred to herein as an improved, adequate, or clinically acceptable X-ray projection image. 【0037】 Returning to the method shown in FIG. 2, in operation S140, a pose metric is output to provide pose information for the x-ray projection image 120. 【0038】 The posture metrics can be output in a variety of ways, including graphically and audibly. For example, the posture metrics can be output graphically on a monitor, tablet, or another device display. As some examples, the posture of the projection X-ray imaging system 140 can be output graphically as an icon similar to that shown at the top of FIG. 6 . Feedback for adjusting the posture of the projection X-ray imaging system can be output in a similar manner by displaying arrows indicating how the posture should be adjusted to obtain improved images of the anatomical structures. Feedback for adjusting the posture of the projection X-ray imaging system can alternatively be output audibly via loudspeakers, headphones, etc. For example, audio messages such as "Rotate the X-ray source 10 degrees to the left" or "Rotate your ankle 5 degrees to the left" can be output. As some other examples, the suitability of the posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120 can be output by outputting a color or another visual indication of whether the posture is proper. For example, the X-ray projection image 120 can have a red outline indicating that the posture is improper and a green outline indicating that the posture is indeed proper. 【0039】 In one embodiment, the posture metric represents feedback for adjusting the posture P of the projection X-ray imaging system to acquire subsequent X-ray projection images representing the anatomical structure 130, and the posture metric is used to automatically adjust the posture P of the projection X-ray imaging system. In this example, the feedback is output to a control system that generates control signals for controlling one or more motors or actuators to adjust the posture of the projection X-ray imaging system that generated the X-ray projection image 120. The projection X-ray imaging system may have various sensors, such as position sensors and rotational encoders, that operate in combination with the one or more motors or actuators to measure the posture of the projection X-ray imaging system and provide the adjusted posture. In this example, the feedback may be output to the user as a recommended posture, which is automatically adjusted in response to the user accepting the recommended posture. This example has the advantage of eliminating the need to manually reposition the projection X-ray imaging system and / or the patient to acquire subsequent X-ray projection images, thereby saving time and improving workflow. 【0040】 In one embodiment, the posture metrics are output and can be stored along with the x-ray projection images 120. The x-ray projection images 120 can be stored in a standardized format, such as the DICOM format. For example, the posture metrics can be stored along with the x-ray projection images 120 in a so-called "secondary capture" DICOM format or a "presentation state" DICOM format. The posture metrics and x-ray projection images 120 in the secondary capture DICOM format or the presentation state DICOM format can then be transmitted to a PACS viewing station for viewing by a radiologist. The posture metrics and x-ray projection images 120 can be displayed on the PACS viewing station, for example, as secondary capture images, presentation state images, or structured reports. 【0041】 In step S130, the projected sub-region 150 1..m The relative sizes of two or more of the controlled sub-regions are used to generate a posture metric for the x-ray projection image. 【0042】 In one approach, the size of the projected subregions is calculated from the X-ray projection images. In this approach, the size of the projected subregions is calculated using image processing techniques. For example, the relative sizes of the projected subregions 1501 and 1502 described above with reference to FIG. 6 can be measured and used as a measure of the rotation of the projection X-ray imaging system about the y-axis of the anatomical structure 130 and, therefore, the pose of the projection X-ray imaging system relative to the anatomical structure. As the rotation of the projection X-ray imaging system is rotated clockwise about the y-axis throughout the sequence of FIGS. 6a-6c, the ratio of the size of the projected subregion 1501 to the size of the projected subregion 1502 decreases. The relative sizes of the other projected subregions can be used to enhance the pose determined from the projected subregions 1501 and 1502, thus improving the accuracy of the determined pose. For example, the ratio of the projected sub-region 1505 to the projected joint gap 1503 between the tibia and talus also changes throughout the sequence of Figures 6a-6c and with rotation about the x-axis, and therefore this can also improve the accuracy of the pose determined from the projected sub-regions 1501 and 1502. 【0043】 As mentioned above, in one embodiment, the pose metric represents the pose P of the projection X-ray imaging system 140 relative to the anatomy 130. This embodiment uses multiple projected sub-regions 150 for the reference anatomy as shown in FIG. 1..mThis can be achieved by measuring the size of the anatomy at multiple poses and generating a functional relationship between the associated sizes and the corresponding poses. The functional relationship can be provided, for example, by a graph or a look-up table. The graph or look-up table can then be queried to determine the pose for a new projection x-ray image representing the anatomical structure. 【0044】 As mentioned above, in another embodiment, the posture metric represents the suitability of a posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120. This embodiment can be implemented in a similar manner by determining a posture P as described above and checking the posture against a range of postures that have been labeled in functional relationships as suitable for acquiring the X-ray projection image 120. In this embodiment, the posture metric represents the suitability of a posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120, and process S130 for generating a posture metric for the X-ray projection image 120 includes determining a posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120 based on two or more projected sub-regions 150. 1..m and comparing the relative sizes of the x-ray projection images 120 to at least one threshold. In this embodiment, the threshold may define a range of poses suitable for acquiring the x-ray projection images 120. A radiologist may set the threshold to define clinically acceptable images for imaging the anatomical structure. 【0045】 As mentioned above, in another embodiment, the posture metric represents feedback for adjusting the posture P of the projection X-ray imaging system for acquiring subsequent X-ray projection images representing the anatomy 130. This embodiment can be realized by calculating the posture for a new projection X-ray image as described above, and calculating the posture adjustment as the difference between the calculated posture and a reference posture for acquiring an X-ray projection image representing the anatomy 130. Referring to FIG. 6 , in this embodiment, the above calculation can provide a posture adjustment that is calculated as the difference between posture values ​​P1 and P2, or as the difference between posture values ​​P3 and P2, where P2 is the optimal posture for the anatomy. 【0046】 In one embodiment, the pose metric for the X-ray projection image 120 is further determined based on anatomical landmarks in the X-ray projection image, which is used to determine the position of the projected sub-region 150 in the segmented X-ray projection image 120. 1..m This embodiment provides redundancy in determining the pose metric based on the relative sizes of two or more projected sub-regions of the x-ray projection image 120. This embodiment is useful for improving the accuracy of the pose metric in process S130 in situations where the x-ray projection image does not allow for accurate segmentation. For example, bone contours may be less visible in some poses and in some x-ray imaging settings. For example, anatomical variations or anatomical implants may obstruct image features in the x-ray projection image and confuse the segmentation process. In this example, the method described with reference to FIG. 1 uses one or more anatomical landmarks 160 in the x-ray projection image 120 to generate a pose metric. 1..n and generating a posture metric for the x-ray projection image 120, S130, comprising identifying the location of one or more anatomical landmarks 160. 1..n Further based on the identified location of the 【0047】 This embodiment is described with reference to FIG. 7, which is an example of a segmented x-ray projection image 120 showing projected sub-regions 150 of an anatomical structure, in accordance with some aspects of the present disclosure. 1..m Examples of multiple anatomical landmarks and 160 1..n 7 includes examples of the positions of the landmarks 1601 and 1602. The segmented X-ray projection image 120 shown in FIG. 7 is an enlarged view of FIG. 5, centered on the tip of the fibula 1601, and includes the positions of the tip of the fibula 1601 and the tip of the tibial malleolus 1602. As can be seen, the landmarks 1601 in the segmented X-ray projection image 120 shown in FIG. 1..n The position of the anatomical landmark changes as the posture of the projection X-ray imaging system changes relative to the ankle shown in Figure 5. For example, as posture changes, the position of a single anatomical landmark in the projection X-ray image 120 changes depending on the position of the projected sub-region 150. 1..mAs a result, the posture metric for the X-ray projection image 120 varies with respect to the position of one or more anatomical landmarks 160 in the X-ray projection image 120. 1..n and the projected sub-region 150 in the segmented X-ray projection image 120. 1..m Similarly, if multiple anatomical landmarks are identified in the projection x-ray image 120, the locations of these landmarks may be further calculated based on the relative sizes of two or more of the projected sub-regions 150. 1..m can be determined for the position of 【0048】 If multiple landmarks are identified in the x-ray projection images 120, other information can be further derived from the x-ray projection images 120 and used to generate posture metrics. For example, in one embodiment, multiple anatomical landmarks 160 1..n 7, in this embodiment, multiple distances 170 between different anatomical landmarks can be measured. 1..i is calculated. For example, distance 170 1..3represents the radii of the circular approximations of the two talar and lower tibial condyles. Distance 1704 represents the distance between the tip of the fibula and the tip of the tibial malleolus. Distance 1705 represents the posterior joint space represented by the distance between the circular approximation of the tibial condyle and the closer talar condyle. Distance 1706 represents the medial joint space represented by the distance between the circular approximation of the tibial condyle and the closer talar condyle. Distance 1708 represents the anterior joint space represented by the distance between the circular approximation of the tibial condyle and the closer talar condyle. Distance 1707 represents the anterior distance between the fibula and tibia at a defined significant position. As can be appreciated, these distances vary according to the orientation of the X-ray imaging system used to generate the X-ray projection image 120 shown in FIG. 7 , and as a result, these distances can be used to characterize the orientation of the X-ray imaging system. In another example, angles between trajectories defined by multiple anatomical landmarks can be measured and used in a manner similar to that used to characterize the pose of an X-ray imaging system used to generate the X-ray projection image 120 shown in FIG. 7. 【0049】 These two examples can be applied to different anatomical structures and can be described more formally as follows: TIFF2025530161000002.tif527 represents the image coordinates, TIFF2025530161000003.tif512 Let denote the grayscale image. A mapping that assigns an anatomical region to every pixel u, TIFF2025530161000004.tif515 is defined, where N is the number of anatomical regions. Furthermore, an additional geometric measure, TIFF2025530161000005.tif611 where L is the number of additional measures. These measures can be distances or radii, as shown in Figure 7. In the example where the pose metric represents feedback for adjusting the projection X-ray imaging system pose, in a second step, these anatomical segments are mapped, TIFF2025530161000006.tif523 where M is the dimension of the feedback for correcting the pose of the projection X-ray imaging system relative to the anatomy. The mapping ξ is a linear or nonlinear mapping from an additional geometric measure to the anatomy and pose feedback. 【0050】 In another embodiment, anatomical landmarks 160 in the X-ray projection image 120 1..n The positions of the anatomical landmarks 160 are mapped to corresponding landmarks in the reference image, and the posture metrics of the X-ray projection images are calculated based on the position of the anatomical landmarks 160 relative to the corresponding landmarks in the reference image. 1..n In this example, the reference image is obtained by projecting a 3D model representing the anatomical structure using different poses of the projection X-ray imaging system relative to the anatomical structure. The 3D model is projected in a virtual sense using a model of the projection X-ray imaging system in which the X-ray source and X-ray detector are positioned in the same manner as the imaging system used to generate the X-ray projection image. The projected landmarks in the reference image appear at positions characteristic of the pose of the model of the projection X-ray imaging system used to generate the reference image. Thus, the anatomical landmarks 160 in the X-ray projection image 120 1..n By matching the positions of the vertices with corresponding landmarks in the reference image, a pose metric can be further determined. 【0051】 Thus, in these embodiments, the method described with reference to FIG. 2 involves generating a plurality of anatomical landmarks 160 in the x-ray projection image 120. 1..n identifying the location of a plurality of anatomical landmarks 160; 1..n one or more distances 170 between the positions of 1..i and / or measuring the angle of one or more trajectories defined by a plurality of anatomical landmarks, and / or measuring the angle of one or more trajectories defined by a plurality of anatomical landmarks 160. 1..nto a plurality of corresponding landmarks in the reference image, and generating a pose metric for the X-ray projection image 120 (S130) includes mapping the position of the measured distance 170 to a plurality of corresponding landmarks in the reference image. 1..i , and / or the measured angles of one or more trajectories, and / or a plurality of anatomical landmarks 160 1..n , each of which is based on a map of the position of the landmark relative to the corresponding landmark in the reference image. 【0052】 In these examples, the anatomical landmarks can be identified using various techniques, such as feature detectors, edge detectors, model-based segmentation, or neural networks, which can be used to identify the anatomical landmarks in the x-ray projection images 120. 【0053】 In the above example, the X-ray projection image 120 is segmented into projected sub-regions 150 within the X-ray projection image 120. 1..m Before determining the relative sizes of two or more of the projected sub-regions 150, the projected sub-regions 150 may be scaled. This reduces the impact on the pose metrics from factors such as differences in object size and differences in the field of view of the projection X-ray imaging system. Thus, in one embodiment, process S130 for generating the pose metrics for the X-ray projection image 120 includes scaling the X-ray projection image 120, and determining the relative sizes of the two or more projected sub-regions 150. 1..m The relative sizes of two or more of the projected sub-regions are determined from the scaled x-ray projection images. 【0054】 In this example, the area of ​​the X-ray projection image 120 can be scaled based on a measurement of the area within the X-ray projection image 120. Alternatively, the area of ​​the X-ray projection image 120 can be scaled based on a measurement of the distance within the X-ray projection image 120. The area of ​​the X-ray projection image can be scaled 120 based on a measurement of the area within the X-ray projection image 120 relative to a reference area. The reference area can be defined, for example, as the projected area of ​​a portion of a bone having a known pose. Similarly, the area of ​​the X-ray projection image 120 can be scaled based on a measurement of the distance within the X-ray projection image 120 relative to a reference distance. The reference distance can be defined, for example, as the length of the bone or the width of the bone at a predetermined location along its length. 【0055】 In another example, an anatomical structure 130 in an X-ray projection image 120 is registered to an anatomical atlas image representing the anatomical structure 130 to provide a scale factor for the X-ray projection image, and the area of ​​the X-ray projection image 120 is scaled using the scale factor. 【0056】 In another approach, the X-ray projection images are input to a neural network, and the neural network is trained to generate a pose metric using a ground truth value of the pose metric calculated based on the relative sizes of two or more of the projected sub-regions in the segmented X-ray projection images. In this approach, the pose metric may similarly represent the pose P of the projection X-ray imaging system 140 relative to the anatomical structure 130 and / or the suitability of the projection X-ray imaging system pose P for acquiring the X-ray projection images 120 and / or feedback P for adjusting the projection X-ray imaging system pose P for acquiring subsequent X-ray projection images representing the anatomical structure 130. 【0057】 In this approach, at S130, generating a posture metric for the X-ray projection image 120 includes inputting segmented image data representing the segmented X-ray projection image into a second neural network NN2, and generating a posture metric using the second neural network NN2 in response to the input, wherein the second neural network NN2 is trained to generate the posture metric from the segmented image data using X-ray projection image training data, the X-ray projection image training data including a plurality of segmented X-ray projection images representing the anatomical structure 130 and corresponding ground truth values ​​of the posture metric, and the ground truth value of the posture metric is calculated based on the relative sizes of two or more of the projected sub-regions in the segmented X-ray projection image. 【0058】 This approach is described with reference to FIG. 8, which is a schematic diagram illustrating an example of a method for providing posture information for an X-ray projection image, in which, according to some aspects of the present disclosure, X-ray projection data 110 including the X-ray projection image is input to a first neural network NN1 to segment the X-ray projection image, and segmented image data representing the segmented image is input to a second neural network NN2 to generate a posture metric for the X-ray projection image. 【0059】 Referring to Figure 8, in this approach, segmented image data representing segmented X-ray projection images is input to a second neural network NN2. In the example shown in Figure 8, the segmented image data is obtained by inputting X-ray projection data 110 to a first neural network NN1. The first neural network NN1 is trained to segment X-ray projection images representing anatomical structures 130. As mentioned above, an example of a neural network that can be used to segment X-ray projection images in process S120 is Kronke, S., et al., "CNN-based pose estimation for assessing quality of ankle-joint X-ray images," Proc. SPIE 12032, Medical Imaging 2022: Image Processing, 120321A (2022). The neural network NN1 can be provided by a variety of architectures, including, for example, a convolutional neural network (CNN), an encoder-decoder based model, a recurrent neural network (RNN) based model, and a generative model with adversarial training (GAN). It should be noted that instead of using a neural network, i.e., NN1, to obtain the segmented image data, the segmented image data can be obtained using a segmentation algorithm such as model-based segmentation, watershed-based segmentation, region growing, level set, or graph cut, as described above. 【0060】 The segmented image data is then input to a second neural network NN2, which can be implemented by a variety of different architectures, including, for example, CNN, ResNet, U-Net, and encoder-decoder architectures. 【0061】 The second neural network NN2 is trained to generate posture metrics from the segmented image data using X-ray projection image training data. The X-ray projection image training data includes a plurality of segmented X-ray projection images representing the anatomical structure 130 and corresponding ground truth values ​​of posture metrics. The X-ray projection image training data may include tens, hundreds, thousands, or more segmented X-ray projection images representing the anatomical structure 130. The X-ray projection image training data may include data from subjects of various ages, body mass index (BMI) values, different genders, and different medical conditions. The segmented X-ray projection images used in the training data may be obtained by segmenting the X-ray projection images. The segmentation may be performed manually by an expert or using various segmentation algorithms, such as those described above. The ground truth values ​​of the posture metrics are calculated based on the relative sizes of two or more of the projected subregions in the segmented X-ray projection images. In this regard, the relative size may be calculated using the image processing techniques described above and used to determine one or more of the following: a pose P of the projection X-ray imaging system 140 relative to the anatomical structure 130; a suitability of the pose P of the projection X-ray imaging system for acquiring the X-ray projection image 120; and feedback for adjusting the pose P of the projection X-ray imaging system for acquiring subsequent X-ray projection images representing the anatomical structure 130. The pose metric may be calculated using the techniques described above. During training, the ground truth value of the pose metric and, optionally, the relative sizes of two or more of the projected sub-regions in the segmented X-ray projection images used to derive the pose metric are input to the neural network. 【0062】 Training a neural network involves inputting a training dataset into the neural network and iteratively adjusting the neural network's parameters until the trained neural network provides accurate outputs. Training is often performed using dedicated neural processors such as graphics processing units (GPUs), neural processing units (NPUs), or tensor processing units (TPUs). Training often employs a centralized approach, in which cloud- or mainframe-based neural processors are used to train the neural network. Following its training with the training dataset, the trained neural network can be deployed to a device where it analyzes new input data during inference. Processing requirements during inference are significantly lower than those needed during training, allowing neural networks to be deployed to a variety of systems, such as laptop computers, tablets, and mobile phones. Inference can be performed, for example, by a central processing unit (CPU), GPU, NPU, or TPU on a server or in the cloud. 【0063】 Therefore, the process of training the above-mentioned neural network NN2 involves adjusting its parameters. The parameters, more specifically, weights and biases, control the processing of activation functions in the neural network. In supervised learning, the training process automatically adjusts the weights and biases so that when presented with input data, the neural network accurately provides the corresponding expected output data. To do this, a loss function or error value is calculated based on the difference between the predicted output data and the expected output data. The value of the loss function can be calculated using functions such as negative log-likelihood loss, mean absolute error (or L1 norm), mean squared error, root mean squared error (or L2 norm), Huber loss, or (binary) cross-entropy loss. During training, the value of the loss function is typically minimized, and training is terminated when the value of the loss function meets a stopping criterion. In some cases, training is terminated when the value of the loss function meets one or more of several criteria. 【0064】 There are many well-known methods for solving loss minimization problems, including gradient descent and quasi-Newton algorithms. A variety of algorithms have been developed to implement these methods and their derivatives, including, but not limited to, stochastic gradient descent (SGD), batch gradient descent, mini-batch gradient descent, Gauss-Newton algorithm, Levenberg-Marquardt algorithm, momentum algorithm, Adam's algorithm, Nadam's algorithm, Ada-Grad's algorithm, Ada-Delta algorithm, RMSProp, and Adamax algorithm. These algorithms use the chain rule to calculate the derivatives of the loss function with respect to the model parameters. This process is called backpropagation because it calculates derivatives from the last layer (output layer) to the first layer (input layer). These derivatives tell the algorithm how to adjust the model parameters to minimize the error function. That is, adjustments to the model parameters begin at the output layer and proceed backward through the network until the input layer is reached. In the first training iteration, the initial weights and biases are often randomized. Next, the neural network predicts output data, which is also random. The weights and biases are then adjusted using backpropagation. The training process is performed iteratively by adjusting the weights and biases at each iteration. Training ends when the error, i.e., the difference between the predicted output data and the expected output data, is within an acceptable range for the training or validation data. The neural network is then deployed, and the trained neural network makes predictions for new input data using the trained parameter values. If the training process is successful, the trained neural network accurately predicts the expected output data from the new input data. 【0065】 As previously mentioned, in one embodiment, the pose metric represents the suitability of the pose P of the projection X-ray imaging system for acquiring the X-ray projection images. In this example, the second neural network NN2 is receiving x-ray projection image training data; inputting the X-ray projection image training data into a second neural network NN2; For each of the segmented X-ray projection images, predicting values ​​of the pose metric using a second neural network NN2; adjusting parameters of a second neural network NN2 based on a difference between the predicted value of the pose metric and the ground truth value of the pose metric; repeating the prediction and adjustment until a stopping criterion is met; and It is trained to generate pose metrics from segmented X-ray projection images by running 【0066】 These processes can be performed according to the backpropagation techniques described above. In this example, the ground truth value of the pose metric, i.e., the goodness of fit, can be calculated automatically or manually by an expert reviewer based on the relative sizes of two or more of the projected subregions in the segmented x-ray projection image. The ground truth value of the pose metric can also be determined based on one or more additional criteria. 【0067】 As mentioned above, in another embodiment, the pose metric represents feedback for adjusting the projection X-ray imaging system pose P to acquire subsequent X-ray projection images representing the anatomical structure 130. In this example, the ground truth value of the pose metric includes one or more pose adjustments for adjusting the pose of the projection X-ray imaging system to acquire X-ray projection images representing the anatomical structure with a target pose relative to the anatomical structure. The second neural network: receiving x-ray projection image training data; inputting the X-ray projection image training data into the second neural network; For each of the segmented X-ray projection images, predicting, using a second neural network, values ​​of a posture metric, the values ​​of the posture metric comprising one or more posture adjustments for adjusting the posture of the projection X-ray imaging system to acquire X-ray projection images representing the anatomical structure at a target posture relative to the anatomical structure; adjusting parameters of the second neural network based on a difference between the predicted value of the pose metric and the ground truth value of the pose metric; repeating the prediction and adjustment until a stopping criterion is met; and It is trained to generate pose metrics from segmented X-ray projection images by running 【0068】 These processes can be performed according to the backpropagation technique described above. In this example, the ground truth value of the pose metric, i.e., feedback, can be determined automatically or manually by an expert reviewer based on the relative sizes of two or more of the projected subregions in the segmented x-ray projection image. The ground truth value of the pose metric can also be determined based on one or more additional criteria. Thus, given the actual size of the projected subregions, it is calculated how to adjust the current pose to acquire a subsequent x-ray projection image representing the anatomical structure 130 in which the relative sizes of the projected subregions are within a predetermined range for the anatomical region. For the subsequent x-ray projection image, this predetermined range can be set so that the resulting x-ray projection image provides a preferable or clinically acceptable image. 【0069】 In another example, the method described above with reference to FIG. 2 also includes generating a value of a second posture metric. The value of the second posture metric is used to adapt the value of the posture metric. In this example, both the posture metric described above with reference to FIG. 2 and the second posture metric represent the suitability of the posture P of the projection X-ray imaging system for acquiring the X-ray projection images 120. However, the second posture metric is calculated based on the size of the gap between two bones in the X-ray projection images or segmented X-ray projection images. The second posture metric is calculated based on different criteria than the posture metric described above with reference to FIG. 1 and can be used to override the assessment provided by the posture metric. 【0070】 In this example, the posture metric represents the suitability of the posture P of the X-ray imaging system for acquiring the X-ray projection image 120, and the method described with reference to FIG. 1 includes a step of calculating a value of a second posture metric of the X-ray projection image 120 based on the size of the gap 1503 between two bones in the X-ray projection image 120, the value of the second posture metric representing the suitability of the posture P of the X-ray imaging system for acquiring the X-ray projection image 120, and the step S130 of generating the posture metric of the X-ray projection image 120 is further based on the value of the second posture metric. 【0071】 In this example, the value of the second posture metric is used to adapt the value of the first posture metric. Here, if the value of the second posture metric indicates that the posture P of the X-ray imaging system is actually suitable for acquiring the X-ray projection image 120 based on the size of the gap 1503 between the two bones in the X-ray projection image 120, the posture metric output in process S140 is adapted from "inappropriate" to "suitable." Thus, if the value of the second posture metric indicates that the posture is suitable for the X-ray projection image, the value overwrites the original evaluation of the posture metric. 【0072】 An example of a gap between two bones in an X-ray projection image 120 is the gap 1503 between the tibia and talus in the X-ray projection image 120 shown in Figure 5. The size of the gap 1503 can be calculated as an area or distance measurement of the gap. The size of the gap can be calculated using the projected sub-region 150 described above. 1..m can be calculated in a similar manner to the size of 【0073】 The value of the second posture metric can be calculated using a method similar to that described above for the posture metric. Thus, in one approach, image processing techniques are used to calculate the size of the gap between two bones in an X-ray projection image. Alternatively, in another approach, described in more detail below, a neural network is trained to generate the value of the second posture metric from X-ray projection data. 【0074】 In an image processing method, image processing techniques can be used to calculate the size of the gap 1503 between two bones in the X-ray projection image 120. Based on this size, the value of the second posture metric can be calculated in various ways. Generally, the larger the size value, the higher the conformance of the posture with respect to the X-ray projection image. Thus, in one example, the size can be used directly as an analog value of the second posture metric. In another example, the size can be digitized based on one or more thresholds and used to generate distinct categories for the value of the second posture metric. In this example, the distinct categories correspond to different levels of conformance. 【0075】 As described above, in another method, a neural network is trained to generate a numerical value of the second posture metric from the X-ray projection data. In this approach, calculating the value of the second posture metric includes inputting X-ray projection data representing the X-ray projection images 120 or segmented X-ray projection images into a third neural network NN3, and generating the value of the second posture metric using the third neural network NN3 in response to the input, where the third neural network NN3 is trained to generate the value of the second posture metric from the X-ray projection data using X-ray projection image training data, the X-ray projection image training data including a plurality of X-ray projection images or segmented X-ray projection images representing the anatomical structure 130, and the corresponding ground truth value of the second posture metric is calculated based on the size of a gap between two bones in the X-ray projection images or segmented X-ray projection images. 【0076】 Referring to the bottom portion of Figure 8, in this example, X-ray projection data representing X-ray projection image 120 or a segmented X-ray projection image is input to a third neural network NN3, as indicated by the dashed line in Figure 8. As described above, the segmented X-ray projection image can be generated using neural network NN1, or alternatively, using various segmentation algorithms. 【0077】 The third neural network NN3 can be implemented by various architectures, including, for example, a convolutional neural network (CNN), ResNet, U-Net, and an encoder-decoder architecture. Examples of neural networks that can be trained for this purpose are disclosed in Mairhofer, D., et al., "An AI-based Framework for Diagnostic Quality Assessment of Ankle Radiographs", Proceedings of Machine Learning Research 143:484-496, 2021. 【0078】 The third neural network NN3 is trained to generate a second posture metric from the X-ray projection data using X-ray projection image training data. The X-ray projection image training data includes a plurality of X-ray projection images or segmented X-ray projection images representing the anatomical structure 130 and corresponding ground truth values ​​of the second posture metric. The X-ray projection image training data may include tens, hundreds, thousands, or more segmented X-ray projection images representing the anatomical structure 130. The X-ray projection image training data may include data from subjects of various ages, body mass index (BMI) values, different genders, and different medical conditions. The ground truth value of the second posture metric is calculated based on the size of a gap between two bones in the X-ray projection images or segmented X-ray projection images. The ground truth value of the second posture metric is calculated by determining the size of a gap between two bones in the X-ray projection images or segmented X-ray projection images using the image processing method described above. In this regard, the size of the gap is calculated and used to determine a value for the suitability of the posture P of the X-ray imaging system for acquiring the X-ray projection images 120. A second posture metric can be calculated for the (segmented) X-ray projection images using the techniques described above. Thus, the size can be used directly as an analog value of the second posture metric, or the size can be digitized based on one or more thresholds and used to generate distinct categories for the values ​​of the second posture metric. During training, the ground truth value of the posture metric, and optionally the relative sizes of two or more of the projected subregions in the segmented X-ray projection images used to derive the posture metric, are input to the third neural network NN3. 【0079】 In this example, the third neural network, NN3, receiving x-ray projection image training data; inputting X-ray projection image training data into a third network NN3; For each of the plurality of X-ray projection images or segmented X-ray projection images, predicting values ​​of a second pose metric using a third neural network NN3; adjusting parameters of a third neural network NN3 based on a difference between the predicted value of the second pose metric and the ground truth value of the second pose metric; repeating the prediction and adjustment until a stopping criterion is met; and The second pose metric is trained to generate values ​​for the second pose metric from the X-ray projection data by running 【0080】 These processes can be performed according to the backpropagation technique described above. 【0081】 In another example, a computer program is provided that, when executed by one or more processors, causes the one or more processors to perform a method for providing pose information for X-ray projection images, the method including the steps of receiving (S110) X-ray projection data 110, the X-ray projection data including X-ray projection images 120 representing an anatomical structure 130, the X-ray projection data having been acquired by a projection X-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130; and receiving (S110) a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m Segmenting the X-ray projection image 120 (S120) to identify projected sub-regions 150 within the segmented X-ray projection image 120. 1..m generating a posture metric for the X-ray projection image 120 based on the relative sizes of two or more of the projected sub-regions (S130); and outputting the posture metric to provide posture information for the X-ray projection image 120 (S140). 【0082】 In another example, a system 200 for providing pose information related to X-ray projection images is provided, the system including one or more processors 210, the one or more processors receiving (S110) X-ray projection data 110, the X-ray projection data including X-ray projection images 120 representing an anatomical structure 130, the X-ray projection data acquired by a projection X-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130; and projecting a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m Segmenting the X-ray projection image 120 (S120) to identify projected sub-regions 150 within the segmented X-ray projection image 120. 1..m and (S140) generating a posture metric for the X-ray projection image 120 based on the relative sizes of two or more of the projected sub-regions. 【0083】 An example of a system 200 is shown in Figure 3. It should be noted that the system 200 may also include one or more of a projection X-ray imaging system 140 for generating X-ray projection data 110, a user interface device (not shown in Figure 3), such as a keyboard, mouse, touch screen, etc., for receiving user input for processing performed by the one or more processors 210, a patient bed (not shown in Figure 3), and a monitor (not shown in Figure 3) for displaying the X-ray projection images 120, output posture metrics, and other data for processing performed by the one or more processors 210. 【0084】 In the above example, a value of a second posture metric was generated and used to adapt the value of the posture metric. In this example, the second posture metric represents the suitability of the posture P of the projection X-ray imaging system for acquiring the X-ray projection image 120, and the second posture metric is calculated based on the size of the gap between two bones in the X-ray projection image or the segmented X-ray projection image. The second posture metric was described as being used to override the evaluation provided by the posture metric. In another example, the second posture metric can simply be output rather than being used to adapt the value of the posture metric. In this example described below, the second posture metric is simply referred to as the posture metric. In this example, a computer-implemented method for providing pose information for an X-ray projection image includes receiving (S110) X-ray projection data 110, the X-ray projection data including X-ray projection images 120 representing an anatomical structure 130, the X-ray projection data having been acquired by a projection X-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130; and receiving (S110) a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m and generating a pose metric for the X-ray projection image 120 (S130), the value of the pose metric representing the suitability of a pose P of the X-ray imaging system for acquiring the X-ray projection image 120, ... 1..m and (S140) generating posture metrics based on sizes of one or more projected sub-regions among the X-ray projection image 120, the sizes including a size of a gap 1503 between two bones in the X-ray projection image 120, and / or a size of a projected sub-region 1501, 1502 that is at least partially defined by a gap 1503 between two bones and another bone in the X-ray projection image 120. 【0085】 This example can be implemented in the same way as described above for the second pose metric. Thus, an example of a gap between two bones in an X-ray projection image 120 according to this embodiment is the gap 1503 between the tibia and talus in the X-ray projection image 120 shown in FIG. 5. The size of the gap 1503 can be calculated as an area or distance measurement of the gap. The size of the gap can be calculated using the projected sub-region 150 described above. 1..m In one embodiment, the size of a gap 1503 between two bones in an X-ray projection image 120 is calculated by measuring the separation between the bones at multiple locations along a portion of the gap, and a posture metric for the X-ray projection image 120 is generated based on the variation in the spacing between the bones along the portion of the gap. The variation in the spacing between the bones along the gap can be measured using various metrics, including, for example, by calculating the variance or standard deviation of the spacing between the bones along the portion of the gap. An example of a gap that can be measured in this manner is the gap 1503 between the tibia and talus in the X-ray projection image 120 shown in FIG. 5. A variation value indicating a relatively high uniformity in the measured separation between the bones along the portion of the gap, e.g., a relatively low variance, can be associated with a proper posture P of the X-ray imaging system acquiring the X-ray projection image 120, while a variation value indicating a relatively low uniformity in the measured separation between the bones along the portion of the gap, e.g., a relatively high variance, can be associated with an improper posture P of the X-ray imaging system acquiring the X-ray projection image 120. This example can be used, for example, to determine posture metrics for anterior-posterior "AP" images of the ankle. 【0086】 Similarly, the value of the pose metric can be generated based on the size of the projected sub-region in the segmented X-ray projection image 120, and the size of the projected sub-region 150 1..mis defined, as described above, at least in part by the intersection of multiple bones (1501, 1502) in the X-ray projection image or by the intersection of a gap 1503 between two bones with another bone in the X-ray projection image 120. The value of the posture metric can be calculated using approaches similar to those described above. Thus, in one approach, image processing techniques are used to calculate the size of the gap between two bones in the X-ray projection image. Alternatively, in another approach, a neural network is trained to generate the value of the posture metric from the X-ray projection data. In an image processing method, image processing techniques can be used to calculate the size of the gap 1503 between two bones in the X-ray projection image 120. The value of the posture metric can be calculated based on this size in various ways. Generally, the larger the size value, the better the posture fit for the X-ray projection image. Thus, in one example, the size can be used directly as an analog value of the posture metric. In another example, the size can be digitized based on one or more thresholds and used to generate discrete categories of posture metric values. In this example, the discrete categories correspond to different levels of suitability. As described above, in another approach, a neural network is trained to generate values ​​of the posture metric from the X-ray projection data. In this approach, calculating the values ​​of the posture metric includes inputting the X-ray projection data representing the X-ray projection image 120 or a segmented X-ray projection image into a third neural network NN3, and generating the values ​​of the posture metric using the third neural network NN3 in response to the input, where the third neural network NN3 is trained to generate the values ​​of the posture metric from the X-ray projection data using X-ray projection image training data, where the X-ray projection image training data includes a plurality of X-ray projection images or segmented X-ray projection images representing the anatomical structure 130 and corresponding ground truth values ​​of the posture metric, where the ground truth value of the posture metric is calculated based on the size of a gap between two bones in the X-ray projection image or segmented X-ray projection image. 【0087】 In this example, the third neural network NN3 is trained to generate values ​​of the posture metric from the X-ray projection data by performing the steps of receiving X-ray projection image training data, inputting the X-ray projection image training data into the third network NN3, predicting, for each of a plurality of X-ray projection images or segmented X-ray projection images, a value of the posture metric using the third neural network NN3, adjusting parameters of the third neural network NN3 based on the difference between the predicted value of the posture metric and a ground truth value of the posture metric, and repeating the prediction and adjustment until a stopping criterion is met. 【0088】 These processes can be performed according to the backpropagation technique described above. 【0089】 A corresponding computer program and a corresponding apparatus are also provided according to this embodiment. 【0090】 The computer program has instructions that, when executed by one or more processors, cause the one or more processors to perform a method for providing pose information for X-ray projection images, the method including the steps of receiving (S110) X-ray projection data 110, the X-ray projection data including X-ray projection images 120 representing an anatomical structure 130, the X-ray projection data having been acquired by a projection X-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130; 1..m and generating a posture metric for the X-ray projection image 120 (S130), the posture metric value representing the suitability of a posture P of the X-ray imaging system for acquiring the X-ray projection image 120, the posture metric value representing the suitability of a posture P of the X-ray imaging system for acquiring the X-ray projection image 120, the posture metric value representing the suitability of a posture P of the X-ray imaging system for acquiring the X-ray projection image 120, the posture metric value representing the suitability of a posture P of the X-ray projection image 120 for the projection sub-region 150. 1..mthe projection images 120 based on the size of one or more of the projected sub-regions 1501, 1502, the sizes including the size of a gap 1503 between two bones in the X-ray projection image 120, and / or the size of a projected sub-region 1501, 1502 defined at least in part by an intersection between a plurality of bones in the X-ray projection image 120, and / or the size of a projected sub-region defined at least in part by an intersection of the gap 1503 between two bones with another bone in the X-ray projection image 120; and outputting a posture metric to provide posture information regarding the X-ray projection image 120 (S140). 【0091】 The corresponding system includes steps of receiving (S110) X-ray projection data 110, the X-ray projection data including X-ray projection images 120 representing an anatomical structure 130, the X-ray projection data being acquired by a projection X-ray imaging system 140 having a corresponding pose P relative to the anatomical structure 130; and generating a plurality of projected sub-regions 150 of the anatomical structure 130. 1..m and generating a posture metric for the X-ray projection image 120 (S130), the value of which represents a suitability of a posture P of an X-ray imaging system that acquires the X-ray projection image 120, and the value of which represents a suitability of a posture P of the projected sub-region 150. 1..m and (S140) outputting a posture metric to provide posture information of the X-ray projection image 120. The X-ray projection image 120 includes one or more processors configured to perform the steps of: generating posture metrics based on sizes of one or more projected sub-regions among the X-ray projection image 120, the sizes including a size of a gap 1503 between two bones in the X-ray projection image 120, and / or sizes of projected sub-regions 1501, 1502 defined at least in part by intersections between multiple bones in the X-ray projection image 120, and / or sizes of projected sub-regions defined at least in part by intersections between the gap 1503 between two bones and other bones in the X-ray projection image 120. 【0092】 In a related example, in S130, the process of generating the posture metric for the X-ray projection image 120 is performed using a normalized value for the size of the gap 1503. The normalized value for the size of the gap 1503 is calculated by i) scaling the size of the gap 1503 or the projected sub-region based on the size of the anatomical feature in the X-ray projection image, or ii) registering the anatomical structure in the X-ray projection image 120 to an anatomical atlas image representing the anatomical structure 130 to provide a scale factor for the X-ray projection image, scaling the area of ​​the X-ray projection image 120 using the scale factor to provide a scaled X-ray projection image, and measuring the size of the gap 1503 or the size of the projected sub-region in the scaled X-ray projection image. 【0093】 In this example, the anatomical features used to generate the normalized value of gap size may be bone dimensions, such as bone length or bone width. Alternatively, an anatomical atlas image may be used. Normalization compensates for differences in feature sizes in the projection images used to determine pose, thereby providing a more reliable pose metric. 【0094】 The following are examples of the present invention. 【0095】 Example 1. 1. A computer-implemented method for providing pose information for an x-ray projection image, comprising: receiving (S110) X-ray projection data (110), the X-ray projection data including X-ray projection images (120) representing an anatomical structure (130), the X-ray projection data being acquired by a projection X-ray imaging system (140) having a corresponding pose (P) relative to the anatomical structure (130); A plurality of projected sub-regions (150) of the anatomical structure (130) 1..m performing segmentation (S120) of the X-ray projection image (120) to identify A projected sub-region (150) in the segmented X-ray projection image (120)1..m generating (S130) a pose metric for the X-ray projection image (120) based on the relative sizes of two or more projected sub-regions of the X-ray projection image (120); outputting a posture metric to provide posture information for the X-ray projection image; A method comprising: 【0096】 Example 2. Projected subregions (150) of anatomical structures (130) 1..m ) is a part of a bone (150 1..2 , 150 4..6 ), and the space between the two bones (1503). 【0097】 Example 3. The anatomical structure (130) comprises a plurality of bones, and at least one of the two or more projected sub-regions is a bone (150) in an X-ray projection image. 1..6 15. The computer-implemented method of claim 1 or 2, wherein the bone is defined at least in part by a perimeter of a portion of at least one bone in the x-ray projection image (120), and / or by intersections of a plurality of bones in the x-ray projection image (120) (1501, 1502), and / or by intersections of a space between two bones in the x-ray projection image (120) with other bones. 【0098】 Example 4. A plurality of projected sub-regions (150) of the anatomical structure (130) 1..m The step of segmenting (S120) the X-ray projection image (120) to identify applying a segmentation algorithm to the received X-ray projection data (110); inputting the received X-ray projection data (110) into a first neural network (NN1), the first neural network (NN1) being trained to segment X-ray projection images representing said anatomical structure (130); 4. The computer-implemented method of any one of Examples 1 to 3, comprising: 【0099】 Example 5. The computer-implemented method of Example 1, wherein the posture metric represents the posture (P) of the projection X-ray imaging system (140) relative to the anatomical structure (130), and / or the posture metric represents the suitability of the posture (P) of the projection X-ray imaging system for acquiring the X-ray projection image (120), and / or the posture metric represents feedback for adjusting the posture (P) of the projection X-ray imaging system for acquiring subsequent X-ray projection images representing the anatomical structure (130). 【0100】 Example 6. The posture metric represents the suitability of the projection X-ray imaging system posture (P) for acquiring the X-ray projection image (120), and the step of generating a posture metric (S130) for the X-ray projection image (120) includes: 1..m 6. The computer-implemented method of any one of Examples 1 to 5, comprising comparing the relative sizes of the plurality of pixels (i.e., the number of pixels) to at least one threshold. 【0101】 Example 7. The method further comprises identifying one or more anatomical landmarks (160) in the X-ray projection image (120). 1..n ), and generating (S130) a posture metric for the X-ray projection image (120) further comprises identifying the location of the one or more anatomical landmarks (160). 1..n 7. The computer-implemented method of any one of Examples 1 to 6, based on the identified location of 【0102】 Example 8. The method further comprises identifying a plurality of anatomical landmarks (160) in the X-ray projection image (120). 1..n ) and the method further comprises: Multiple anatomical landmarks (160 1..n ) one or more distances between the positions (170 1..i), and / or measuring the angle of one or more trajectories defined by a plurality of anatomical landmarks, and / or measuring the angle of one or more trajectories defined by a plurality of anatomical landmarks (160 1..n ) to a plurality of corresponding landmarks in a reference image; The generation of pose metrics (S130) for the X-ray projection images (120) may be performed using one or more measured distances (170 1..i ), and / or a plurality of anatomical landmarks (160) of one or more trajectories and / or corresponding landmarks in the reference image. 1..n 8. The computer-implemented method of example 7, further based on the measured angles of the mapped positions of 【0103】 Example 9. The step of generating (S130) a pose metric for the X-ray projection image (120) includes scaling the X-ray projection image (120) to obtain a pose metric for the projected sub-region (150). 1..m 9. The computer-implemented method of any one of Examples 1 to 8, wherein the relative sizes of two or more projected sub-regions of the scaled x-ray projection image are determined from the scaled x-ray projection image. 【0104】 Example 10. 10. The computer-implemented method of Example 9, wherein scaling the X-ray projection image includes scaling an area of ​​the X-ray projection image (120) based on an area measurement within the X-ray projection image (120) or scaling an area of ​​the X-ray projection image (120) based on a distance measurement within the X-ray projection image (120). 【0105】 Example 11. The step (S130) of generating a posture metric for the X-ray projection image (120) comprises: inputting the segmented image data representing the segmented X-ray projection images into a second neural network (NN2); generating a pose metric using a second neural network (NN2) in response to the input; and and wherein a second neural network (NN2) is trained to generate a posture metric from the segmented image data using X-ray projection image training data, the X-ray projection image training data comprising a plurality of segmented X-ray projection images (130) representing anatomical structures and corresponding ground truth values ​​of the posture metric, the ground truth values ​​of the posture metric being calculated based on relative sizes of two or more of the projected sub-regions in the segmented X-ray projection images. 【0106】 Example 12. The pose metric represents the suitability of the pose (P) of the projection X-ray imaging system for acquiring the X-ray projection images, and the second neural network (NN2) receiving x-ray projection image training data; inputting the X-ray projection image training data into a second neural network (NN2); For each of the segmented X-ray projection images, predicting values ​​of the pose metric using a second neural network (NN2); adjusting parameters of a second neural network (NN2) based on a difference between the predicted value of the pose metric and the ground truth value of the pose metric; repeating the prediction and adjustment until a stopping criterion is met; and 12. The computer-implemented method of Example 11, wherein the computer-implemented method is trained to generate pose metrics from segmented X-ray projection images by performing 【0107】 Example 13. the pose metric represents feedback for adjusting the projection X-ray imaging system pose (P) for acquiring subsequent X-ray projection images representing the anatomical structure (130), and the ground truth value of the pose metric comprises one or more pose adjustments for adjusting the pose of the projection X-ray imaging system having a target pose relative to the anatomical structure for acquiring X-ray projection images representing the anatomical structure; The second neural network is receiving x-ray projection image training data; inputting the X-ray projection image training data into the second neural network; For each of the segmented X-ray projection images, predicting, using a second neural network, values ​​of a posture metric, the values ​​of the posture metric comprising one or more posture adjustments for adjusting the posture of the projection X-ray imaging system to acquire X-ray projection images representing the anatomical structure at a target posture relative to the anatomical structure; adjusting parameters of a second neural network based on a difference between the predicted value of the pose metric and the ground truth value of the pose metric; and repeating the prediction and adjustment until a stopping criterion is met. and 12. The computer-implemented method of Example 11, wherein the computer-implemented method is trained to generate pose metrics from segmented X-ray projection images by performing 【0108】 Example 14. The posture metric represents the suitability of the X-ray imaging system posture (P) for acquiring the X-ray projection image (120), the method further comprising: calculating a value of a second posture metric for the X-ray projection image (120) based on a size of a gap (1503) between two bones in the X-ray projection image (120), the value of the second posture metric representing a suitability of a posture (P) of the X-ray imaging system for acquiring the X-ray projection image (120); 10. The computer-implemented method of Example 1, wherein generating (S130) a posture metric for the X-ray projection image is further based on values ​​of a second posture metric. 【0109】 Example 15. The step of calculating a value of a second pose metric comprises: inputting the X-ray projection image (120) or X-ray projection data representing the segmented X-ray projection image into a third neural network (NN3); generating a value of a second pose metric using a third neural network (NN3) in response to the input; and wherein a third neural network (NN3) is trained to generate values ​​of the second posture metric from the x-ray projection data using x-ray projection image training data, the x-ray projection image training data comprising a plurality of x-ray projection images or segmented x-ray projection images representing the anatomical structure (130) and corresponding ground truth values ​​of the second posture metric, the ground truth values ​​of the second posture metric being calculated based on a size of a gap between two bones in the x-ray projection images or segmented x-ray projection images. 【0110】 The above examples should be understood to illustrate, but not limit, the present disclosure. Further examples are contemplated. For example, an example described in connection with a computer-implemented method may also be provided in a corresponding manner by a computer program, a computer-readable storage medium, or an X-ray imaging system. It should be understood that features described with respect to any one example may be used alone or in combination with other described features, and may be used in combination with one or more other features of the example or with combinations of other examples. Furthermore, equivalents and modifications not described above may be employed without departing from the scope of the present invention as defined in the appended claims. In the claims, the word "comprising" does not exclude other elements or operations, and the indefinite articles "a" or "an" do not exclude a plurality. The mere fact that certain features are recited in mutually different dependent claims does not indicate that a combination of these features cannot be advantageously used. Any reference signs in the claims should not be construed as limiting their scope.

Claims

[Claim 1] A computer implementation method for providing orientation information related to an X-ray projection image, A step of receiving X-ray projection data, wherein the X-ray projection data has an X-ray projection image representing an anatomical structure having multiple bones, and the X-ray projection data was acquired by a projection X-ray imaging system having a posture corresponding to the anatomical structure. The steps include segmenting the X-ray projection image to identify multiple projected subregions of the anatomical structure, A step of generating a posture metric for an X-ray projection image based on the relative sizes of two or more projected subregions among the plurality of projected subregions in the segmented X-ray projection image, wherein the posture metric represents the posture of the projection X-ray imaging system with respect to the anatomical structure, and / or the suitability of the posture of the projection X-ray imaging system for acquiring the X-ray projection image, and / or feedback for adjusting the posture of the projection X-ray imaging system to acquire subsequent X-ray projection images representing the anatomical structure. At least one of the projected subregions represents the space between the two bones, and / or At least one of the projected subregions is at least partially defined by the gap between the two bones and the intersection with the other bone in the X-ray projection image. Steps and The steps include outputting the attitude metric and providing the attitude information of the X-ray projection image, A method of having. [Claim 2] The method according to claim 1, wherein at least one of the projected subregions represents a bone portion. [Claim 3] The method according to claim 1, wherein at least one of the other projected subregions is at least partially defined by the periphery of at least one bone portion of the plurality of bones in the X-ray projection image. [Claim 4] The step of segmenting the X-ray projection image to identify multiple projected subregions of the anatomical structure is, The step of applying a segmentation algorithm to the received X-ray projection data, or A step of inputting the received X-ray projection data into a first neural network, wherein the first neural network is trained to segment the X-ray projection images representing the anatomical structures, The method according to claim 1, comprising: [Claim 5] The aforementioned attitude metric represents the suitability of the attitude of the projection X-ray imaging system that acquires the X-ray projection image. The method according to claim 1, wherein the step of generating the attitude metric of the X-ray projection image includes comparing the relative sizes of the two or more projected subregions with at least one threshold. [Claim 6] The above method further, Step 1: Identifying the location of one or more anatomical landmarks in the X-ray projection image. It has, The method according to claim 1, wherein the step of generating the posture metric of the X-ray projection image is further based on the identified positions of one or more anatomical landmarks. [Claim 7] The method comprises the step of identifying the locations of a plurality of anatomical landmarks in the X-ray projection image, and the method further comprises the step of A step of measuring one or more distances between the locations of the plurality of anatomical landmarks, and / or A step of measuring the angle of one or more trajectories defined by the plurality of anatomical landmarks, and / or The steps include mapping the locations of the aforementioned multiple anatomical landmarks to a plurality of corresponding landmarks in a reference image, It has, The method according to claim 6, wherein the step of generating the posture metric of the X-ray projection image is based on the measured distances and / or the measured angles of the one or more trajectories and / or the positions of the plurality of anatomical landmarks mapped to the corresponding landmarks in the reference image. [Claim 8] The step of generating the attitude metric of the X-ray projection image includes scaling the X-ray projection image, The method according to claim 1, wherein the relative sizes of two or more projected subregions among the projected subregions are determined from the scaled X-ray projection image. [Claim 9] The scaling of the aforementioned X-ray projection image is Scaling the area of ​​the X-ray projection image based on the measured area within the X-ray projection image, or Scaling the area of ​​the X-ray projection image based on the measured distance within the X-ray projection image, or The anatomical structures in the X-ray projection image are registered to an anatomical atlas representing the anatomical structures to provide a scale factor for the X-ray projection image, and the area of ​​the X-ray projection image is scaled using the scale factor. The method according to claim 8, wherein the method is characterized by having the following: [Claim 10] The step of generating the attitude metric of the X-ray projection image is, The steps include inputting the segmented image data representing the segmented X-ray projection image into a second neural network, The steps include generating the attitude metric using the second neural network in response to the input, The second neural network is trained to generate the posture metric from the segmented image data using X-ray projection image training data, the X-ray projection image training data comprising a plurality of segmented X-ray projection images representing the anatomical structure and the corresponding ground truth value of the posture metric, the ground truth value of the posture metric is calculated based on the relative sizes of two or more projected subregions among the projected subregions in the segmented X-ray projection image. The method according to claim 1. [Claim 11] The attitude metric represents the suitability of the attitude of the projection X-ray imaging system for acquiring the X-ray projection image, and the second neural network is The steps include receiving X-ray projection image training data, The steps include inputting the X-ray projection image training data into the second neural network, For each of the segmented X-ray projection images, The steps include: predicting the value of the posture metric using the second neural network described above; The steps include adjusting the parameters of the second neural network based on the difference between the predicted value of the attitude metric and the ground truth value of the attitude metric, The process involves repeating the prediction step and the adjustment step until the stopping criteria are met. The steps to perform, The method according to claim 10, wherein the system is trained to generate the pose metric from the segmentation X-ray projection image by performing the following: [Claim 12] The posture metric represents feedback for adjusting the posture of the projection X-ray imaging system to acquire a subsequent X-ray projection image representing the anatomical structure, and the ground truth value of the posture metric includes one or more posture adjustments for adjusting the posture of the projection X-ray imaging system to acquire an X-ray projection image representing the anatomical structure having a target posture relative to the anatomical structure. The second neural network described above is: The steps include receiving X-ray projection image training data, The steps include inputting the X-ray projection image training data into the second neural network, For each of the segmented X-ray projection images, A step of predicting the value of the posture metric using the second neural network, wherein the value of the posture metric has one or more posture adjustments for adjusting the posture of the projection X-ray imaging system to obtain an X-ray projection image representing the anatomical structure having a target posture relative to the anatomical structure. The steps include adjusting the parameters of the second neural network based on the difference between the predicted value of the attitude metric and the ground truth value of the attitude metric, The process involves repeating the prediction step and the adjustment step until the stopping criteria are met. The steps to perform, The method according to claim 10, wherein the system is trained to generate the pose metric from the segmented X-ray projection image by performing the following: [Claim 13] The attitude metric represents the suitability of the attitude of the X-ray imaging system that acquires the X-ray projection image, and the method further, A step of calculating a second posture metric value for the X-ray projection image based on the size of the gap between two bones in the X-ray projection image, wherein the value of the second posture metric represents the suitability of the posture of the X-ray imaging system for acquiring the X-ray projection image. The method according to claim 1, wherein the step of generating the attitude metric of the X-ray projection image is further based on the value of the second attitude metric. [Claim 14] The step of calculating the value of the second attitude metric is, The steps include inputting the X-ray projection image or the X-ray projection data representing the segmented X-ray projection image into a third neural network, The steps include generating the value of the second posture metric using the third neural network in response to the input, The method according to claim 13, wherein the third neural network is trained to generate a value for the second posture metric from the X-ray projection data using the X-ray projection image training data, the X-ray projection image training data comprising a plurality of X-ray projection images or segmented X-ray projection images representing the anatomical structure and a corresponding ground truth value for the second posture metric, the ground truth value for the second posture metric is calculated based on the size of the gap between two bones in the X-ray projection image or segmented X-ray projection image. [Claim 15] A computer implementation method for providing orientation information related to an X-ray projection image, A step of receiving X-ray projection data, wherein the X-ray projection data has an X-ray projection image representing an anatomical structure, and the X-ray projection data was acquired by a projection X-ray imaging system in a position corresponding to the anatomical structure. The steps include segmenting the X-ray projection image to identify multiple projected subregions of the anatomical structure, A step of generating an attitude metric for the X-ray projection image, wherein the value of the attitude metric represents the suitability of the attitude of the X-ray imaging system that acquires the X-ray projection image, and the value of the attitude metric is generated based on the size of one or more projected subregions among the projected subregions, and the size is The size of the gap between the two bones in the aforementioned X-ray projection image, and / or The size of the projected subregion, at least partially defined by the gap between the two bones and the intersection with the other bone in the aforementioned X-ray projection image, Steps including, The steps include outputting the attitude metric and providing attitude information of the X-ray projection image, A method of having. [Claim 16] The step of generating the posture metric of the X-ray projection image is performed using a normalized value of the size of the gap between two bones in the X-ray projection image, and the normalized value of the size of the gap is, i) Scaling the size of the gap or the projected sub-region based on the size of the features in the X-ray projection image, or ii) Registering the anatomical structures in the X-ray projection image against an anatomical atlas image representing the anatomical structures to provide a scale factor for the X-ray projection image, scaling the area of ​​the X-ray projection image using the scale factor to provide the scaled X-ray projection image, and measuring the size of the gap or the size of the projected sub-region within the scaled X-ray projection image. The method according to claim 15, calculated by...

Images