Bone mineral density acquisition method, X-ray imaging system and storage medium

Abstract

The invention provides a bone mineral density acquisition method, an X-ray imaging system and a storage medium. The bone mineral density obtaining method comprises the steps that at least one X-ray image of a detected object is obtained through an X-ray imaging system, the at least one X-ray image is an original image or a medical image obtained after image processing, the at least one X-ray image is processed on the basis of a trained learning network, and the bone mineral density is obtained. The result comprises at least one of a T value of bone mineral density and classification, and at least one of an abnormal bone position, an abnormal probability and an abnormal prompt, and the abnormality means that the T value exceeds a threshold range.

Description

Technical Field

[0001] This invention relates to medical imaging technology, and more specifically to a method for obtaining bone density based on X-ray images, an X-ray imaging system, and a non-transitory computer-readable storage medium. Background Technology

[0002] In an X-ray imaging system, radiation from an X-ray source is directed at a subject, typically a patient in a medical diagnostic application. A portion of the radiation passes through the subject and impacts a detector, which is divided into a matrix of discrete elements (e.g., pixels). The detector elements are read out to generate an output signal based on the amount or intensity of radiation impacting each pixel region. The signal can then be processed to produce a medical image that can be displayed for examination on the display device of the X-ray imaging system.

[0003] Typically, to obtain the bone mineral density of a subject, it is necessary to use a professional bone mineral density meter to measure it and obtain the patient's T-value. However, most subjects nowadays need to undergo regular physical examinations, and taking chest X-rays is a necessary part of these examinations. Therefore, it is desirable to obtain the bone mineral density of the subject based on the obtained chest X-rays or X-ray images, so that the subject does not have to undergo unnecessary radiation exposure. Summary of the Invention

[0004] This invention provides a method for obtaining bone density based on X-ray images, an X-ray imaging system, and a non-transitory computer-readable storage medium.

[0005] An exemplary embodiment of the present invention provides a method for obtaining bone density based on X-ray images. The method includes acquiring at least one X-ray image of a subject using an X-ray imaging system, wherein the at least one X-ray image is a raw image or a medical image after image processing; and a trained learning network to output a result of the bone density of the subject, and at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of a T-value and a classification of the bone density, wherein the abnormality refers to the result exceeding a threshold range.

[0006] An exemplary embodiment of the present invention also provides a non-transitory computer-readable storage medium for storing a computer program that, when executed by a computer, causes the computer to perform the above-described bone density acquisition method.

[0007] An exemplary embodiment of the present invention also provides an X-ray imaging system. The X-ray imaging system includes a control device capable of performing the bone density acquisition method described above.

[0008] An exemplary embodiment of the present invention also provides an X-ray imaging system. The X-ray imaging system includes an acquisition unit and an image processing unit. The acquisition unit is capable of acquiring at least one X-ray image of a subject under test using the X-ray imaging system. The at least one X-ray image is either a raw image or a medical image after image processing. The image processing unit is capable of outputting a bone density result of the subject under test, and at least one of abnormal bone location, abnormal probability, and abnormal indication, based on a trained learning network. The result includes at least one of a T-value and a classification of the bone density, and the abnormality refers to the result exceeding a threshold range.

[0009] Other features and aspects will become clear from the following detailed description, accompanying drawings, and claims. Attached Figure Description

[0010] The invention can be better understood by describing exemplary embodiments of the invention in conjunction with the accompanying drawings, in which:

[0011] Figure 1 This is a schematic diagram of an X-ray imaging system according to some embodiments of the present invention;

[0012] Figure 2 This is a schematic diagram of an X-ray imaging system according to other embodiments of the present invention;

[0013] Figure 3 This is a schematic diagram of a control device according to some embodiments of the present invention;

[0014] Figure 4 This is a schematic diagram of a control device according to other embodiments of the present invention;

[0015] Figure 5 This is a schematic diagram of a control device according to some embodiments of the present invention;

[0016] Figure 6 This is a schematic diagram of a learning network according to some embodiments of the present invention;

[0017] Figure 7 This is a schematic diagram of an X-ray image with indicator markings according to some embodiments of the present invention;

[0018] Figure 8 This is a schematic diagram of an X-ray image with indicator markings according to other embodiments of the present invention;

[0019] Figure 9 This is a flowchart of a bone density acquisition method according to some embodiments of the present invention; and

[0020] Figure 10This is a flowchart of a method for obtaining bone density according to other embodiments of the present invention. Detailed Implementation

[0021] The following describes specific embodiments of the present invention. It should be noted that, in order to provide a concise description, this specification cannot exhaustively describe all features of the actual embodiments. It should be understood that, in the actual implementation of any embodiment, just as in any engineering or design project, various specific decisions are often made to achieve the developer's specific goals and to meet system-related or business-related constraints, and this can change from one embodiment to another. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this invention, some design, manufacturing, or production modifications based on the technical content disclosed herein are merely conventional technical means and should not be construed as insufficient content of this disclosure.

[0022] Unless otherwise defined, the technical or scientific terms used in the claims and description shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in the patent application description and claims of this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. The terms "an" or "a" and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms "comprising" or "including" and similar terms mean that the element or object preceding "comprising" or "including" encompasses the element or object listed following "comprising" or "including" and its equivalents, and do not exclude other elements or objects. The terms "connected" or "linked" and similar terms are not limited to physical or mechanical connections, nor are they limited to direct or indirect connections.

[0023] Figure 1 An X-ray imaging system 100 according to some embodiments of the present invention is shown. Figure 2 An X-ray imaging system 200 according to other embodiments of the present invention is shown. For example... Figure 1 As shown, the X-ray imaging system 100 includes a suspension device 110, a wall stand device 120, and a test bed device 130. The suspension device 110 includes a longitudinal guide rail 111, a transverse guide rail 112, a telescopic cylinder 113, a trolley 114, an X-ray tube assembly 115, and an X-ray tube control device 116.

[0024] Although this application is based on Figure 1The suspended X-ray imaging system shown is used as an example for description; however, the bone density acquisition method and apparatus in this application can also be applied to... Figure 2 In the ground-rail X-ray imaging system and / or the mobile X-ray imaging system shown, specifically, the X-ray source can be mounted on a ground-rail crossarm, that is, the crossarm that mounts the X-ray source is mounted on a track on the ground via a column, and the X-ray source can move along the track, column and crossarm. Of course, the X-ray source can also be mounted on a mobile trolley via a telescopic arm.

[0025] For ease of description, in this application, the x-axis, y-axis and z-axis are defined as follows: the x-axis and y-axis are located in the horizontal plane and are perpendicular to each other, and the z-axis is perpendicular to the horizontal plane. Specifically, the direction of the longitudinal guide rail 111 is defined as the x-axis, the direction of the transverse guide rail 112 is defined as the y-axis, and the extension direction of the telescopic cylinder 113 is defined as the z-axis, which is the vertical direction.

[0026] The longitudinal guide rail 111 and the transverse guide rail 112 are arranged vertically, wherein the longitudinal guide rail 111 is mounted on the ceiling, and the transverse guide rail 112 is mounted on the longitudinal guide rail 111. The telescopic cylinder 113 is used to carry the X-ray tube assembly 115.

[0027] A trolley 114 is positioned between the transverse guide rail 112 and the telescopic cylinder 113. The trolley 114 may include a rotating shaft, a motor, and a drum. The motor drives the drum to rotate around the rotating shaft, thereby causing the telescopic cylinder 113 to move along the z-axis and / or slide relative to the transverse guide rail. The trolley 114 can slide relative to the transverse guide rail 112, meaning it can drive the telescopic cylinder 113 and / or the X-ray tube assembly 115 to move along the y-axis. Furthermore, the transverse guide rail 112 can slide relative to the longitudinal guide rail 111, thereby causing the telescopic cylinder 113 and / or the X-ray tube assembly 115 to move along the x-axis.

[0028] The telescopic cylinder 113 includes multiple cylindrical sections with different inner diameters, which can be sequentially fitted into the upper cylinder from bottom to top to achieve telescopic movement. The telescopic cylinder 113 is telescopic (or movable) in the vertical direction, that is, the telescopic cylinder 113 can drive the X-ray tube assembly 115 to move along the z-axis. The lower end of the telescopic cylinder 113 is also provided with a rotating part, which can drive the X-ray tube assembly 115 to rotate.

[0029] Specifically, the X-ray source and collimator 117 are housed within the X-ray tube assembly 115, with the collimator 117 typically mounted below the X-ray source. The size of the collimator 117 opening determines the X-ray irradiation range, i.e., the size of the field of view (FOV). X-rays can pass through the collimator opening and irradiate the region of interest (ROI) of the object being inspected, while other X-rays are absorbed by the leaves to prevent the object from absorbing excessive and unnecessary doses.

[0030] In some embodiments, the X-ray imaging system 100 further includes a camera unit 140 aligned with the detector for acquiring real-time camera images of the object being inspected. Additionally, the camera can also acquire images of the detector, etc. Specifically, the camera unit 140 is mounted on the suspension device 110, and more specifically, on the side of the collimator 117.

[0031] Camera unit 140 may include one or more cameras, such as digital cameras, analog cameras, depth cameras, infrared cameras, or ultraviolet cameras, or 3D cameras, 3D scanners, or red-green-blue (RGB) sensors, RGB-D depth sensors, or other devices capable of capturing color image data of a target object. In some embodiments, camera unit 140 is further provided with a control module capable of controlling the rotation of the camera unit to adjust the shooting range of the camera unit. In other embodiments, the camera unit is a panoramic camera, capable of capturing images of the entire body of the object being detected.

[0032] The camera unit 140 can acquire depth information or depth images of the object being detected. Typically, the depth information is obtained through 3D point cloud computing acquired by the camera. Furthermore, the real-time optical images can be used to acquire at least one of the following: thickness, height, position, body position, and pose of the object being detected.

[0033] In some embodiments, the camera unit 140 may be a camera unit installed in a fixed position within the scanning chamber, or fixed in any other manner. In some embodiments, the optical image acquired by the camera unit 140 is not limited to a single optical image, but may also include a dynamic real-time video stream, i.e., a series of real-time optical images.

[0034] The X-ray tube control device (console) 116 is mounted on the X-ray tube assembly 115. The X-ray tube control device 116 includes a user interface such as a display screen and control buttons for pre-image preparation, such as patient selection, protocol selection, and positioning.

[0035] The movement of the suspension device 110 includes the movement of the X-ray tube assembly along the x-axis, y-axis, and z-axis, as well as the rotation of the X-ray tube assembly in the horizontal plane (rotation axis parallel to or coincident with the z-axis) and the vertical plane (rotation axis parallel to the y-axis). In these movements, a motor typically drives the rotating shaft to rotate the corresponding components, thereby achieving the corresponding movement or rotation. The corresponding control components are generally installed within the trolley 114. The X-ray imaging system further includes a motion control device (not shown in the figure) capable of controlling the aforementioned movements of the suspension device 110. Furthermore, the motion control device can receive control signals to control the corresponding components to perform corresponding movements, thereby driving the X-ray tube assembly to a preset or designated position.

[0036] like Figure 2 As shown, the X-ray imaging system 200 includes a floor unit 210, a column unit 220, and a testing bed unit 230. The floor unit 210 includes a support column 211, a cantilever 212, and an X-ray tube assembly 215. The cantilever 212 supports the X-ray tube assembly 215 and is mounted on the support column 211.

[0037] For ease of description, in this application, the x-axis, y-axis, and z-axis are defined as follows: the x-axis and y-axis are located in the horizontal plane and are perpendicular to each other, and the z-axis is perpendicular to the horizontal plane. Specifically, the extension direction of the cantilever 212 or the width direction of the testing bed device is defined as the x-axis, the direction in the horizontal plane that is perpendicular to the extension direction of the cantilever or the length direction of the testing bed device is defined as the y-axis, and the extension direction of the support column 211 is defined as the z-axis. The z-axis is the vertical direction.

[0038] The floor-mounted device 210 further includes a guide rail mounted on the floor, which is arranged along the y-axis, and the support column 211 moves along the guide rail, i.e., moves along the y-axis. Furthermore, the cantilever 212 is also capable of moving relative to the support column 211 in a vertical direction (i.e., the z-axis direction). Additionally, a drive device may be provided between the X-ray tube assemblies 215, which can drive the X-ray tube assemblies 215 to rotate about the x-axis as a central axis.

[0039] The X-ray tube assembly 215 includes an X-ray source, a collimator 217, and a tube control device 216. The collimator 217 and the tube control device 216 are connected to... Figure 1 The collimator 117 and the X-ray tube control device 116 have roughly similar structures and functions.

[0040] like Figures 1 to 2 As shown, the column assembly 120 / 220 includes a first detector 121 / 221, a column 122 / 222, and a connecting part 123. Figure 2(Not shown). The connecting part 123 includes a support arm perpendicularly connected to the height direction of the columns 122 / 222 and a rotating bracket mounted on the support arm. The first detector 121 / 221 is mounted on the rotating bracket. The column assembly 120 / 220 further includes a detector driving device disposed between the rotating bracket and the first detector 121 / 221. Driven by the detector driving device, the detector moves in a direction parallel to the height direction of the columns 122 / 222 on the plane supported by the rotating bracket. The first detector 121 / 221 can also rotate relative to the support arm, forming a certain angle with the columns. The first detector 121 / 221 has a plate-like structure with a variable orientation, so that the X-ray incident surface can be made vertical or horizontal according to the incident direction of the X-rays.

[0041] The examination bed device 130 / 230 includes a second detector 131 / 231. The selection or use of the first detector 121 / 221 and the second detector 131 / 231 can be determined based on the patient's imaging site and / or imaging protocol, or based on the position of the subject being examined obtained by the camera, so as to perform imaging examinations in a supine or standing position.

[0042] For ease of display, Figure 2 The display unit and other components located in the control room are omitted; however, those skilled in the art should understand that, for example... Figure 2 The X-ray imaging system shown also includes a similar structure. The X-ray imaging system 100 / 200 further includes a display unit 150, which is operatively connected to the camera unit. The display unit 150 includes a user interface 151 for displaying real-time optical images, X-ray images, medical images, information about the object being examined, an exposure parameter setting interface, an image post-processing interface, and so on.

[0043] Specifically, the display unit 150 can include any type of display screen, such as the main display screen located in the control room, the display screen of the X-ray tube control device 116 / 216 located in the scanning room, or a movable display, such as a tablet computer or mobile phone.

[0044] The X-ray imaging system 100 / 200 further includes an input unit 160 for receiving user operations. The input unit 160 may include a touch screen, keyboard, mouse, voice-activated control device, or any other suitable input device. The user can input operation signals / control signals to the control device through the input unit 160.

[0045] The X-ray imaging system 100 / 200 further includes a control device (not shown), which may be a main control device located in a control room, a tube control device, a portable or mobile control device, or any combination thereof. The control device may include a source control device and a detector control device. The source control device is used to command the X-ray source to emit X-rays for image exposure. The detector control device is used to select a suitable detector from among multiple detectors and to coordinate the control of various detector functions, such as automatically selecting the corresponding detector based on the position or posture of the object being inspected, or performing various signal processing and filtering functions, specifically for initial adjustment of dynamic range, interleaving of digital image data, etc. In some embodiments, the control device may provide power and timing signals for controlling the operation of the X-ray source and detector.

[0046] In some embodiments, the control device may also be configured to use digital signals to reconstruct one or more desired images and / or determine useful diagnostic information corresponding to the patient, wherein the control device may include one or more dedicated processors, graphics processing units, digital signal processors, microcomputers, microcontrollers, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other suitable processing devices.

[0047] Of course, the X-ray imaging system 100 / 200 may also include other numbers, configurations, or forms of control devices. For example, the control devices may be local (e.g., located in the same location as one or more X-ray imaging systems 100, such as within the same facility and / or the same local network); in other implementations, the control devices may be remote and therefore accessible only via a remote connection (e.g., via the Internet or other available remote access technologies). In certain implementations, the control devices may also be configured in a cloud-like manner and may be accessed and / or used in a manner substantially similar to accessing and using other cloud-based systems.

[0048] In some embodiments, the X-ray imaging system 100 / 200 further includes an operator workstation that allows a user to receive and evaluate reconstructed images, as well as input control commands (operation signals or control signals). The operator workstation may include a user interface (or user input device), such as a keyboard, mouse, voice-activated control device, or any other suitable input device, through which the operator can input operation signals / control signals to the control device.

[0049] In order to determine or predict bone density based on the acquired X-ray images, Figure 3 A schematic diagram of a control device 300 according to some embodiments of this application is shown. For example... Figure 3As shown, the control device 300 includes an acquisition unit 310 and an image processing unit 320.

[0050] The acquisition unit 310 is capable of acquiring at least one X-ray image of the object being detected, wherein the at least one X-ray image is either the original image or a medical image after image processing.

[0051] The image processing unit 320 includes a learning network and is capable of processing at least one X-ray image based on the trained learning network to output a result of bone density of the detected object, as well as at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of bone density T-value and classification, wherein the abnormality refers to the result exceeding a threshold range.

[0052] Specifically, the acquisition unit 310 can be connected to an X-ray source and a detector to control the X-ray source to emit X-rays toward the object being detected. The detector can detect the output signal attenuated by the object being detected and can acquire the raw image based on the output signal. This raw image, after image processing, becomes a medical image. Both the raw image and the medical image can be used as X-ray images. Specifically, the image processing here is not limited to post-image processing, such as image reconstruction, smoothing, denoising, and artifact removal, but can also include other types of image processing, such as image stitching or tomographic synthesis.

[0053] Figure 6 A schematic diagram of a learning network 600 according to some embodiments of this application is shown, such as... Figure 6 As shown, the X-ray image includes at least one or a combination of single X-ray images, dual-energy X-ray images, stitched images, and TOMO tomography images, or at least one of the above combined with a low-dose local image.

[0054] In some embodiments, the input to the learning network can be a single raw image or a single medical image, such as a chest X-ray, or it can be two raw images obtained with different exposure intensities or medical images obtained through dual-energy operation, or at least two raw images obtained through an image stitching protocol or a stitched image (i.e., a medical image) obtained through image stitching processing. Of course, it can also be multiple raw images obtained through TOMO tomography, or three-dimensional medical images obtained through processing. All of these images are obtained by exposure under the original or existing scanning protocol of the subject being tested. That is to say, even if bone density prediction or calculation is not required, these images still need to be obtained, which itself does not increase the radiation exposure for the subject being tested.

[0055] Of course, in other embodiments, for certain specific subjects, such as older subjects, doctors or users may choose to take an additional low-dose local image under the original scanning protocol. This local image is for a specific area of ​​the subject, such as a bone-dense area or an area prone to osteoporosis or fracture. Specifically, this local image could be the lumbar spine area. By combining the image obtained under the original scanning protocol with the low-dose image of the lumbar spine, the bone density of the subject can be better predicted or calculated. By inputting the combined image from the original scanning protocol and the low-dose local image into the learning network, the obtained bone density results can be more accurate, and the subject does not need to be exposed excessively.

[0056] In some embodiments, the image processing unit 320 includes a classification unit 322 and a judgment unit 323.

[0057] Specifically, classification unit 322 performs classification processing on at least one X-ray image to output a bone density classification. Judgment unit 323 can further output or acquire at least one of abnormal bone location, abnormal probability, and abnormal indication, wherein the abnormality refers to the T value exceeding a threshold range.

[0058] Specifically, bone mineral density (BMD) is an important indicator of bone strength, expressed in grams per cubic centimeter (g / cm³), and is an absolute value. In clinical practice, because different bone mineral density analyzers produce different absolute values, the T-score is typically used to determine whether bone mineral density is normal. The T-score is a relative value, with a normal reference range between -1 and +1.

[0059] Bone mineral density (BMD) classification refers to the classification based on the T-value range of BMD. For example, it can be divided into three categories, including Category I, Category II, and Category III, etc. These classifications are based on the classification methods and criteria known to doctors, users, or those generally skilled in the art. Of course, it can also be divided into fewer or more categories.

[0060] In some embodiments, the T-value of the output bone mineral density is a specific numerical value. The bone mineral density classification can be the first, second, and third categories in Chinese, or the level 1, level 2, and level 3 in English. Of course, it can also be any other suitable representation, such as outputting words like normal, slightly abnormal, or severely abnormal. This application does not limit this.

[0061] Figure 4 A schematic diagram of a control device 400 according to other embodiments of this application is shown, and... Figure 3 The control device shown is different, Figure 4The image processing unit 420 in the control device shown includes a regression unit 424 and a judgment unit 323. The regression unit 424 is capable of performing regression processing on at least one X-ray image region to output the T-value of bone mineral density.

[0062] Figure 5 A schematic diagram of a control device 300 according to some embodiments of this application is shown. Figure 3 The control device shown is different, Figure 5 The image processing unit 520 in the control device shown includes a classification unit 322, a regression unit 424, and a judgment unit 323.

[0063] Therefore, the image processing unit, or learning network, can be customized according to the user's needs. For example, if the user only wants to see the bone density T-value of the detected object, a regression unit needs to be set in the image processing unit. If the user only wants to see the bone density classification of the detected object, a classification unit needs to be set in the image processing unit. If the user wants to see both the bone density T-value and the classification at the same time, both a classification unit and a regression unit need to be set at the same time.

[0064] In some embodiments, both the classification unit and the regression unit can be learning networks, which can be trained based on training data. When the region of interest of the detected object is input into the learning network, the learning network can output the corresponding T value and / or classification.

[0065] Specifically, for the classification unit, multiple sets of X-ray images of the tested object and their corresponding bone density classifications are used as training data. The X-ray images serve as known input, and the bone density classifications are the desired output. This allows the learning network to learn. Once trained, the X-ray image to be classified can be input into the learning network to obtain the corresponding bone density classification. Similarly, the regression unit follows a similar principle: X-ray images are used as known input, and the bone density T-value is the desired output. This allows the learning network to learn. Once trained, the X-ray image to be classified can be input into the learning network to obtain the corresponding bone density T-value. The learning network will be described in detail later.

[0066] Please return to the reference. Figure 3The judgment unit 323 can output or acquire at least one of abnormal bone location, abnormal probability, and abnormal indication. Abnormality refers to a bone density result exceeding a threshold range. Specifically, this could be a T-value exceeding a set threshold or a bone density classification exceeding a set limit. Specifically, when a bone density result exceeds a threshold range, it indicates a possible or high probability of osteoporosis. Different threshold ranges can be set to indicate different types or degrees of abnormality. In some embodiments, different threshold ranges can be set based on the information of the tested subject, such as age or gender. Specifically, a lower threshold range is set for older tested subjects, and a higher threshold range is set for younger tested subjects. Customized thresholds allow for better output of different abnormal locations, probabilities, and / or indications.

[0067] In some embodiments, abnormal bone locations can be displayed on a graphical user interface using images or text with location markers. Specifically, the determination unit or image processing unit can output an X-ray image with indicator markers, wherein the indicator markers on the X-ray image indicate the location of bones that may be osteoporotic. Figure 7 The diagram illustrates X-ray images with indicator marks from some embodiments of this application, such as... Figure 7 As shown, the X-ray image 710 can be any one of at least one images input into the image processing unit, and at least one indicator mark is superimposed on the X-ray image 710, for example, including a first mark 701 located at the scapula position and a second mark 702 located at the lumbar spine position, wherein the at least one indicator mark can highlight different bone positions in any form, for example, the bone position can be marked with an indicator box, or the bone position can be filled with color, or the bone position can be displayed with a mask, etc.

[0068] In some embodiments, when there are multiple abnormal bone locations, there may also be multiple indicator markers, and the judgment unit can use different types or colors of markers to indicate different degrees of abnormality. Specifically, with Figure 7 For example, when the degree of osteoporosis in the scapula is high, a highly saturated color can be used to indicate the location of the scapula; when the degree of abnormality in the lumbar spine is low, a less saturated color or different colors can be used to indicate the specific location of the lumbar spine. Figure 7 This is merely an example and does not limit the scope of this application.

[0069] In some embodiments, except Figure 7 In addition to the X-ray images with indicator markers shown, other types of indicator images can also be output, for example, Figure 8 Schematic diagrams of X-ray images 800 with indicator marks from other embodiments of this application are shown, such as... Figure 8 As shown, abnormal bone locations can be displayed on a graphical user interface using infrared images, or different degrees of abnormality can be displayed using different colors or different color saturations. Of course, abnormal bone locations can also be displayed using other types of heat maps.

[0070] In other embodiments, the determination unit may also indicate the location of a bone by outputting the name or marker of the bone with an abnormal bone position, such as outputting "the fourth joint of the lumbar vertebra" or "L4", etc.

[0071] In some embodiments, the anomaly probability can be an overall probability calculated for all bones in the X-ray image of the object being detected, or it can be a separate probability calculated for the location of the abnormal bone. This anomaly probability can be displayed in the same manner or form as the abnormal bone location, or it can be displayed separately to the user, for example, when the bone location is displayed as... Figure 7 or Figure 8 If the image is presented in a way that indicates the probability, the probability can be indicated by text labels near the bone location, or it can be indicated at a fixed location in the image. When the bone location is indicated by the bone name or bone label, the probability can be indicated after the bone name, for example, left scapula, 60%, etc.

[0072] In some embodiments, an abnormality alert refers to the ability of the judgment unit or image processing unit to issue a prompt to the user or doctor in various ways, such as text or sound, when the T-value or classification of the bone density of the detected object exceeds the threshold range. For example, displaying a pop-up window or prompt icon on the graphical user interface, or issuing a warning sound, etc.

[0073] In some embodiments, the image processing unit 320 further includes a recognition unit 321. Specifically, the recognition unit 321 is capable of recognizing at least one region of interest (ROI) in the X-ray image, and the classification unit 322 performs classification processing on the at least one ROI to output a bone density classification. Furthermore, the recognition unit 321 can also be further used to segment the at least one ROI. Specifically, the at least one ROI may include the bone of the object being detected; that is, the recognition unit is capable of recognizing and / or segmenting the bone portion of the object being detected.

[0074] Specifically, the image processing unit 320 further includes an adjustment unit 325, which adjusts the at least one region of interest based on user input and performs classification processing based on the adjusted region of interest. By setting the adjustment unit 325, users can easily adjust the region of interest manually or assign different weights to different regions of interest. For example, only the bone density of the lumbar spine can be considered, or the weight of the lumbar spine can be greater than that of the scapula, which can be more beneficial for subsequent classification or regression processing, etc.

[0075] Although the control device is divided into multiple unit modules in the above embodiments, those skilled in the art should understand that these multiple unit modules can also be integrated together. That is, the control device of the X-ray imaging system in some embodiments of this application can be used to perform: acquiring at least one X-ray image of the object being tested, wherein the at least one X-ray image is a raw image or a medical image after image processing; processing the at least one X-ray image based on a trained learning network to output the bone density result of the object being tested, and at least one of abnormal bone location, abnormal probability, and abnormal indication, wherein the result includes at least one of the T value and classification of the bone density, and the abnormality refers to the result exceeding a threshold range.

[0076] Specifically, the bone density results, abnormal bone locations, abnormal probabilities, and / or abnormal indications mentioned above are all results generated or output by a computer. These results are merely probability values ​​and are only intended to provide doctors with a reference for accurately diagnosing diseases and developing treatment plans.

[0077] In some embodiments, the functions of the identification unit 321, the classification unit 322 (and / or the regression unit) and the judgment unit 323 are all implemented through a learning network. Of course, the identification unit and the classification unit can be configured to be implemented by a learning network, while the judgment unit can be a separate module integrated into the control device. Of course, the identification unit can also be a separately configured module.

[0078] Further reference Figure 6 When at least one X-ray image is input into the learning network 600, the learning network can output the T-value and / or classification of bone mineral density, as well as the location and / or probability and / or indication of abnormal bone.

[0079] In some embodiments, information about the object to be detected is input into a learning network 600, and the at least one X-ray image is processed based on the information about the object to be detected. The information about the object to be detected includes at least one of the following: the object's age, gender, weight, and exposed body part.

[0080] In some embodiments, the learning network in this application may include any suitable neural network model, which is trained using a clinical dataset. The training dataset includes clinical data of multiple subjects being tested, and the clinical data of each subject includes an X-ray image and its corresponding bone mineral density T-value.

[0081] The learning network may include an input layer, an output layer, and a processing layer (or hidden layer). The input layer is used to preprocess the input data or image, such as removing the mean, normalizing, or reducing dimensionality. The processing layer may include a convolutional layer for feature extraction, a batch normalization layer for standard normal distribution of the input, and an activation layer for nonlinear mapping of the output of the convolutional layer. In addition, it includes a fully connected layer for outputting the corresponding offset.

[0082] Each convolutional layer includes a number of neurons, and the number of neurons in each layer can be the same or different as needed. Based on a first or third dataset (known input) and a second or fourth dataset (desired output), the mathematical relationship between the known input and the desired output is identified and / or the mathematical relationship between the input and output of each layer is identified and characterized by setting the number of processing layers in the network and the number of neurons in each processing layer, and by estimating (or adjusting or calibrating) the weights and / or biases of the network.

[0083] Specifically, when the number of neurons in one layer is n, and the corresponding values ​​of these n neurons are X1, X2...X... n The number of neurons in the next layer connected to one of the layers is m, and the corresponding values ​​of these m neurons are Y1, Y2...Y... m Then the relationship between the two adjacent layers can be represented as:

[0084]

[0085] Among them, X i Y represents the value corresponding to the i-th neuron in the previous layer. j W represents the value corresponding to the j-th neuron in the next layer. ji B represents the weight. j This represents the deviation. In some embodiments, the function f is a corrected linear function.

[0086] Therefore, by adjusting the weight W ji and / or deviation B j This allows us to identify the mathematical relationship between the input and output of each layer, enabling the loss function to converge and thus train the model described above.

[0087] In one embodiment, while the configuration of the learning network is guided by prior knowledge of the estimation problem, dimensions of the inputs, outputs, etc., it relies on or specifically on achieving the best approximation of the desired output data based on the input data. In various alternative implementations, certain aspects and / or features of the data, imaging geometry, reconstruction algorithms, etc., can be leveraged to give explicit meaning to certain data representations in the deep learning network, which can help accelerate training. This is because it creates the opportunity to train (or pre-train) or define certain layers separately within the learning network.

[0088] As discussed in this paper, deep learning techniques (also known as deep machine learning, hierarchical learning, or deep structured learning) employ artificial neural networks for learning. Deep learning methods are characterized by using one or more network architectures to extract or model data of interest. Deep learning methods can be accomplished using one or more processing layers (e.g., input layers, output layers, convolutional layers, normalization layers, sampling layers, etc., with varying numbers and functions depending on the deep network model). The configuration and number of layers allow deep networks to handle complex information extraction and modeling tasks. Specific parameters (also called “weights” or “biases”) of the network are typically estimated through a so-called learning process (or training process). The parameters learned or trained usually result in (or output) a network corresponding to different levels of layers; therefore, extracting or modeling different aspects of the initial data or the output of a previous layer can often represent the hierarchical structure or cascade of layers. In image processing or reconstruction, this can be characterized as different layers relative to different levels of features in the data. Thus, processing can be layered; that is, earlier or higher-level layers may correspond to extracting “simple” features from the input data, followed by layers that combine these simple features into features exhibiting higher complexity. In practice, each layer (or more specifically, each "neuron" within a layer) can employ one or more linear and / or nonlinear transformations (so-called activation functions) to process the input data into an output data representation. The number of "neurons" can be constant across multiple layers or can vary from layer to layer. As discussed in this paper, as part of the initial training of a deep learning process to solve a specific problem, the training dataset includes known input values ​​(e.g., sample images or pixel matrices of images after coordinate transformation) and the expected (target) output value of the final output of the deep learning process (e.g., an image or recognition judgment result). In this way, the deep learning algorithm can process this training dataset (in a supervised or guided manner or in an unsupervised or unguided manner) until it identifies the mathematical relationship between the known inputs and the expected output and / or identifies and characterizes the mathematical relationship between the inputs and outputs of each layer. The learning process typically utilizes (partial) input data and creates a network output for that input data, then compares the created network output with the expected output of the dataset, and then uses the difference between the created and expected outputs to iteratively update the network parameters (weights and / or biases). The parameters of a network can typically be updated using the stochastic gradient descent (SGD) method; however, those skilled in the art will understand that other methods known in the art can also be used to update the network parameters.Similarly, a separate validation dataset can be used to validate the trained network, where both the known input and the desired output are known. The network output can be obtained by feeding the known input to the trained network, and then the network output is compared with the (known) desired output to validate previous training and / or prevent overtraining.

[0089] In some embodiments, the network trained described above is trained on a training module on an external carrier (e.g., a device other than a medical imaging system). In some embodiments, the training system may include a first module for storing the training dataset, a second module for training and / or updating the model, and a communication network for connecting the first and second modules. In some embodiments, the first module includes a first processing unit and a first storage unit, wherein the first storage unit stores the training dataset, and the first processing unit receives relevant instructions (e.g., retrieve training dataset) and sends the training dataset according to the instructions. Furthermore, the second module includes a second processing unit and a second storage unit, wherein the second storage unit stores the training model, and the second processing unit receives relevant instructions, trains and / or updates the network, etc. In other embodiments, the training dataset may also be stored in the second storage unit of the second module, and the training system may not include the first module. In some embodiments, the communication network may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0090] Once data (e.g., a trained model) is generated and / or configured, it can be copied and / or loaded into the X-ray imaging system 100 / 200, which can be done in various ways. For example, the model can be loaded via a directional connection or link between the X-ray imaging system 100 / 200 and the control unit. In this regard, communication between different components can be accomplished using available wired and / or wireless connections and / or according to any suitable communication (and / or network) standards or protocols. Alternatively or additionally, data can be loaded into the X-ray imaging system 100 / 200 indirectly. For example, data can be stored on a suitable machine-readable medium (e.g., a flash memory card, etc.) and then loaded into the X-ray imaging system 100 / 200 (on-site, such as by the system's user or authorized personnel) using that medium, or data can be downloaded to an electronic device capable of local communication (e.g., a laptop computer, etc.) and then used on-site (e.g., by the system's user or authorized personnel) to upload the data to the X-ray imaging system 100 / 200 via a direct connection (e.g., a USB connector, etc.).

[0091] Figure 9 This is a flowchart of a bone density acquisition method based on X-ray images according to some embodiments of the present invention. Figure 9As shown, the bone density acquisition method 900 includes steps 910 and 920.

[0092] In step 910, at least one X-ray image of the object being examined is acquired using an X-ray imaging system, wherein the at least one X-ray image is either the original image or a medical image after image processing.

[0093] Specifically, the X-ray image includes at least one or a combination of a single X-ray image, a dual-energy X-ray image, a stitched image, and a TOMO tomographic image, or at least one of the above combined with a low-dose local image. Specifically, the image obtained in step 910 can be a single raw X-ray image, or a medical image after image post-processing. It can also be two raw images obtained with different exposure doses, or a medical image after dual-energy processing. Furthermore, it can be the original image of multiple sub-X-ray images to be stitched together, or a stitched image, etc.

[0094] In step 920, the trained learning network processes the at least one X-ray image to output the bone density result of the detected object, as well as at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of bone density T-value and classification, and abnormality is defined as the result exceeding a threshold range.

[0095] Specifically, step 920 further includes classifying the at least one X-ray image to output a bone density classification; and / or performing regression processing on the at least one X-ray image to output a bone density T-value.

[0096] In some embodiments, the T-value of the output bone mineral density is a specific numerical value. The bone mineral density classification can be the first, second, and third categories in Chinese, or the level 1, level 2, and level 3 in English, or any other suitable representation.

[0097] An abnormality refers to a bone mineral density (BMD) result exceeding a threshold range. Specifically, this could be a T-score exceeding a set threshold, or a BMD classification exceeding a set limit. In essence, when a BMD result exceeds a threshold range, it indicates a potential risk of osteoporosis or a higher probability of having osteoporosis. Different threshold ranges can be set to output different types of abnormalities or degrees of abnormality. In some embodiments, different threshold ranges can be set based on the information of the subject being tested, such as age or gender.

[0098] The output of abnormal bone locations includes images or text with location markers. Specifically, this includes outputting X-ray images, infrared images, or other types of thermal images with indicator marks, or the names or markers of the bones with abnormal locations. When multiple abnormal bone locations are present, multiple indicator marks can also be used, and different types, colors, or color saturations can be used to identify or display them to indicate different degrees of abnormality. This image or text can be displayed through the graphical user interface of the display unit, or it can be displayed together with the bone density results of the tested object.

[0099] The anomaly probability can be an overall probability calculated for all bones in the X-ray image of the object being examined, or it can be a separate probability calculated for the location of an abnormal bone. This anomaly probability can be shown in the same manner or form as the abnormal bone location, or it can be displayed separately to the user. In some preferred embodiments, the anomaly probability is typically displayed simultaneously on the image with location markers, for example, near the marked bone, to indicate the anomaly probability at that location.

[0100] Anomaly alerts refer to various prompts, such as text or sound, given to users or doctors when the T-value or classification of the bone mineral density of the tested object exceeds the threshold range.

[0101] Specifically, step 920 further includes inputting the information of the object to be detected into the trained learning network, and processing the at least one X-ray image based on the information of the object to be detected. Specifically, the information of the object to be detected includes at least one of the object's age, gender, weight, and exposed body part.

[0102] Figure 10 This is a flowchart of a bone density acquisition method based on X-ray images according to other embodiments of the present invention. Figure 10 As shown, the bone density acquisition method 1000 includes steps 1010, 1020, 1030 and 1040.

[0103] In step 1010, at least one X-ray image of the object being examined is acquired using an X-ray imaging system, wherein the at least one X-ray image is either the original image or a medical image after image processing.

[0104] In step 1020, at least one region of interest in the X-ray image is identified.

[0105] Specifically, at least one region of interest may include the bones of the object being detected, meaning that the recognition unit is able to identify and / or segment the bone portion of the object being detected.

[0106] In some embodiments, step 1020 further includes segmenting the at least one region of interest. In some embodiments, step 1020 further includes adjusting the at least one region of interest based on user input.

[0107] In step 1030, the at least one region of interest is classified to output a bone mineral density classification; and / or the at least one region of interest is regressed to output a bone mineral density T-value.

[0108] In step 1040, based on the bone mineral density classification or T-value, at least one of the following is output: abnormal bone location, abnormal probability, and abnormal indication.

[0109] The bone density acquisition method provided by this invention has several advantages. First, it can measure or acquire bone density based on X-ray images that have already been captured. On the one hand, it does not require the use of a specific bone densitometer for detection, and on the other hand, it does not require the subject to receive additional X-ray exposure. Second, by setting up a learning network, it can acquire the corresponding bone density classification or T-value, as well as various possible abnormal locations, probabilities, or prompts, providing users with intuitive and clear prompts.

[0110] An exemplary embodiment of the present invention provides a method for obtaining bone density based on X-ray images. The method includes acquiring at least one X-ray image of a subject using an X-ray imaging system, wherein the at least one X-ray image is a raw image or a medical image after image processing; and a trained learning network to output a result of the bone density of the subject, and at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of a T-value and a classification of the bone density, wherein the abnormality refers to the result exceeding a threshold range.

[0111] Specifically, the X-ray image includes at least one or a combination of single X-ray images, dual-energy X-ray images, stitched images, and TOMO tomography images, or at least one of the above combined with a low-dose local image.

[0112] Specifically, the bone density acquisition method further includes inputting the information of the object to be detected into the trained learning network, and processing the at least one X-ray image based on the information of the object to be detected.

[0113] Specifically, the information of the object being tested includes at least one of the following: the object's age, gender, weight, and exposed body part.

[0114] Specifically, the learning network based on training processes the at least one X-ray image to classify the at least one X-ray image to output a bone density classification; and / or to regress the at least one X-ray image to output a bone density T-value.

[0115] Specifically, processing the at least one X-ray image further includes identifying at least one region of interest in the at least one X-ray image and classifying and / or regressing the at least one region of interest.

[0116] Specifically, processing the at least one X-ray image further includes: adjusting the at least one region of interest based on user input, and performing classification and / or regression processing based on the adjusted region of interest.

[0117] Specifically, outputting the abnormal bone location includes outputting an X-ray image with a location marker, wherein the marker indicates the abnormal bone location on the X-ray image.

[0118] Specifically, the X-ray images with location markers include different types or colors of markers to indicate different degrees of abnormality.

[0119] An exemplary embodiment of the present invention also provides an X-ray imaging system. The X-ray imaging system includes a control device capable of performing the bone density acquisition method described above.

[0120] An exemplary embodiment of the present invention also provides an X-ray imaging system. The X-ray imaging system includes an acquisition unit and an image processing unit. The acquisition unit is capable of acquiring at least one X-ray image of a subject under test using the X-ray imaging system. The at least one X-ray image is either a raw image or a medical image after image processing. The image processing unit is capable of outputting a bone density result of the subject under test, and at least one of abnormal bone location, abnormal probability, and abnormal indication, based on a trained learning network. The result includes at least one of a T-value and a classification of the bone density, and the abnormality refers to the result exceeding a threshold range.

[0121] Specifically, the X-ray image includes at least one or a combination of single X-ray images, dual-energy X-ray images, stitched images, and TOMO tomography images, or at least one of the above combined with a low-dose local image.

[0122] Specifically, the learning network further processes at least one X-ray image based on information about the object being detected. Specifically, the information about the object being detected includes at least one of the following: age, gender, weight, and exposed body part.

[0123] Specifically, the image processing unit includes a classification unit and / or a regression unit, as well as a judgment unit. The classification unit is used to classify the at least one X-ray image to output a bone density classification. The regression unit is used to perform regression processing on the at least one X-ray image to output a bone density T-value. The judgment unit is used to output at least one of abnormal bone location, abnormal probability, and abnormal indication, wherein the abnormality refers to the result exceeding a threshold range.

[0124] Specifically, the image processing unit further includes a recognition unit for identifying at least one region of interest (ROI) in the at least one X-ray image, and the classification unit and / or regression unit perform classification and / or regression processing on the at least one ROI. Specifically, the image processing unit further includes an adjustment unit for adjusting the at least one ROI based on user input.

[0125] Specifically, the judgment unit can output an X-ray image with location markers, wherein the indicator marks on the X-ray image indicate the location of abnormal bones. Specifically, the X-ray image with location markers includes different types or colors of markers to indicate different degrees of abnormality.

[0126] The present invention may also provide a non-transitory computer-readable storage medium for storing an instruction set and / or a computer program that, when executed by a computer, causes the computer to perform the image processing allocation method described above. The computer executing the instruction set and / or the computer program may be a computer of a medical imaging system or other devices / modules of a medical imaging system. In one embodiment, the instruction set and / or the computer program may be programmed into the processor / control device of the computer.

[0127] Specifically, when this instruction set and / or computer program is executed by the computer, it causes the computer to:

[0128] Acquire at least one X-ray image of the object being inspected, wherein the at least one X-ray image is either the original image or a medical image after image processing;

[0129] The trained learning network processes the at least one X-ray image to output the bone density result of the detected object, as well as at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of the T value of the bone density and classification, and abnormality is defined as the result exceeding a threshold range.

[0130] As described above, instructions can be combined into a single instruction for execution, or any instruction can be split into multiple instructions for execution. Furthermore, the execution order of instructions is not limited to that described above.

[0131] As used herein, the term "computer" can include any processor-based or microprocessor-based system, including systems that use microcontrollers, reduced instruction set computers (RISCs), application-specific integrated circuits (ASICs), logic circuits, and any other circuitry or processors capable of performing the functions described herein. The examples above are merely illustrative and are not intended to limit the definition and / or meaning of the term "computer" in any way.

[0132] Some exemplary embodiments have been described above; however, it should be understood that various modifications can be made. For example, suitable results may be achieved if the described techniques are performed in a different order and / or if components in the described system, architecture, device, or circuit are combined in a different manner and / or replaced or supplemented by other components or their equivalents. Accordingly, other embodiments also fall within the scope of the claims.

Claims

1. A method for obtaining bone density based on X-ray images, comprising: At least one X-ray image of the object being examined is acquired using an X-ray imaging system, wherein the at least one X-ray image is either a raw image or a medical image after image processing; and The trained learning network processes the at least one X-ray image to output the bone density result of the detected object, as well as at least one of abnormal bone location, abnormal probability, and abnormal indication. The result includes at least one of the T value of the bone density and the classification. The abnormality refers to the result exceeding a threshold range.

2. The bone density acquisition method as described in claim 1, wherein, The X-ray images include at least one or a combination of single X-ray images, dual-energy X-ray images, stitched images, and TOMO tomography images, or at least one of the above combined with low-dose local images.

3. The bone density acquisition method as described in claim 1, wherein, It further includes: inputting the information of the object to be detected into the trained learning network, and processing the at least one X-ray image based on the information of the object to be detected.

4. The bone density acquisition method as described in claim 3, wherein, The information of the object being tested includes at least one of the following: the object's age, gender, weight, and exposed body part.

5. The bone density acquisition method as described in claim 1, wherein, Processing the at least one X-ray image based on a trained learning network includes: The at least one X-ray image is classified to output a bone density classification; and / or the at least one X-ray image is regressed to output a bone density T-value.

6. The bone density acquisition method as described in claim 5, wherein, Processing the at least one X-ray image further includes: Identify at least one region of interest in the at least one X-ray image, and classify and / or regress the at least one region of interest.

7. The bone density acquisition method as described in claim 6, wherein, Processing the at least one X-ray image further includes: adjusting the at least one region of interest based on user input, and performing classification and / or regression processing based on the adjusted region of interest.

8. The bone density acquisition method as described in claim 1, wherein, Outputting the abnormal bone location includes outputting an X-ray image with a location identifier, wherein the location identifier marks the abnormal bone location on the X-ray image.

9. The bone density acquisition method as described in claim 8, wherein, The X-ray images with location markers include different types or colors of markers to indicate different degrees of abnormality.

10. A non-transitory computer-readable storage medium for storing a computer program that, when executed by a computer, causes the computer to perform the bone density acquisition method according to any one of claims 1 to 9.

11. An X-ray imaging system, comprising: A control device capable of performing the bone density acquisition method according to any one of claims 1 to 9.

12. An X-ray imaging system, comprising: An acquisition unit is configured to acquire at least one X-ray image of the object being examined using an X-ray imaging system, wherein the at least one X-ray image is a raw image or a medical image after image processing; and An image acquisition unit is used to process the at least one X-ray image based on a trained learning network to output a result of bone density of the detected object, and at least one of abnormal bone location, abnormal probability and abnormal indication, the result including at least one of T value of bone density and classification, the abnormality being that the result exceeds a threshold range.

Images