Detection of anatomical abnormalities in 2D medical images

Abstract

The present invention relates to detecting anatomical abnormalities in medical images. A computer-implemented method 100 and system are disclosed that detects 2D contours 130 of anatomical features in medical images and compares these contours to predicted 2D contours 140 based on a 3D reference model to detect anatomical abnormalities 150. This approach may improve the accuracy of anatomical abnormality detection, thereby reducing time in medical facilities and potentially improving operator experience and patient outcomes.

Description

Technical Field

[0001] The present invention relates to a computer-implemented method, a computer program product, and a system configured to detect anatomical abnormalities in 2D medical images.

Background Art

[0002] Medical images are often used for diagnostic purposes. For example, medical images are examined by a human observer for anatomical abnormalities. Examples of such abnormalities include fractures and tumor tissues that can be identified in X-rays, ultrasound, CT (Computed Tomography), or other medical images.

[0003] However, the examination of medical images by a human observer is time-consuming and error-prone depending on the expertise, experience, time constraints, and fatigue of the human observer. This is especially the case in the harsh environment of a busy clinic and when anatomical abnormalities are not easily distinguishable.

[0004] U.S. Patent Application Publication No. 2022 / 0044041 discloses an algorithm that provides guidance on methods for detecting bone fractures, classifying fractures, and treating fractures.

Summary of the Invention

Problems to be Solved by the Invention

[0005] In particular, an object of the present invention is to provide computer-implemented anatomical abnormality detection in medical imaging.

Means for Solving the Problems

[0006] The present invention is defined by the independent claims. Advantageous embodiments are defined in the dependent claims.

[0007] One aspect of the present invention provides a computer-implemented method for detecting anatomical abnormalities in 2D medical images. The method comprises - Steps to acquire 2D medical images, - A step of detecting the 2D contour of anatomical features, - Steps to obtain a 3D model representing anatomical features, - A step of predicting the 2D contours of anatomical features based on a 3D model, - A step of detecting anatomical abnormalities based on the detected 2D contour and the predicted 2D contour, This approach allows for the detection of anatomical abnormalities. In this way, 2D contours detected in X-ray images can be compared to known reference contours obtained from 3D models to detect and identify differences that may indicate anatomical abnormalities. This approach can improve the accuracy of anatomical abnormality detection, thereby reducing time in healthcare facilities and potentially improving operator experience and patient outcomes.

[0008] The step of predicting the 2D contours of anatomical features based on a 3D model is: - A step of estimating the posture of anatomical features from 2D medical images, - Steps to adjust the 3D model to the estimated pose, - A step of generating a 2D projection from a posture-adjusted 3D model, - A step of predicting the 2D contour of anatomical features from 2D projection, In other words, based on the detected 2D contour, the observation position and posture of the imaged object can be estimated. The 3D model can then be rotated and translated to match the position of the imaged anatomical feature relative to the virtual imaging system, and the 3D model can be adjusted by rotating its own joints, for example by changing the flexion of the ankle joint, to match the posture of the imaged anatomical feature. This creates a healthy reference model with the same field of view and posture parameters as the imaged object, which can be used to predict the 2D reference contour.

[0009] The step of detecting 2D contours may further include a step of segmenting the 2D medical image. For example, individual pixels of the acquired medical image may be labeled as "bone / not bone," for instance, to identify bones in the image and thereby detect 2D contours. Alternatively, the step of detecting 2D contours may include an end-to-end trained machine learning algorithm. In addition, the step of detecting 2D contours may include identifying anatomical features. For example, identifying that the 2D contour is the contour of one or more bones in an ankle joint or wrist. This approach may be further used to optimize the workflow for any further processing or to provide guidance to the user or machine for further steps.

[0010] The method may further include classifying anatomical abnormalities. The step of classifying anatomical abnormalities may be further performed by a machine learning algorithm. Thus, the output of the computer-implemented method is not binary, for example, whether or not an anatomical abnormality was detected, but provides the user insight into the classification of the abnormality. For example, whether or not it is severe, and / or whether or not it is a particular type of fracture and / or ligament rupture.

[0011] According to embodiments of the present invention, the 3D model is -2D medical images, - Anatomical features, and / or - User input It is obtained based on this.

[0012] The 3D model is further, -Computer-aided design models, - A reference model constructed from 3D medical images of anatomical features in at least one posture, Any one of them is acceptable.

[0013] In some cases, the method is -Detected 2D contours and - Predicted 2D contour and The step of generating a difference between is further included.

[0014] Furthermore, a difference map may be generated. The difference map includes information suitable for display on a display.

[0015] In some examples, the anatomical feature is - bone tissue, and - ligament tissue any one of them.

[0016] In some examples, the anatomical abnormality is - fracture, and - ligament rupture any one of them.

[0017] In some examples, the step of classifying the anatomical abnormality is - Weber classification, - Pauwels classification, - Smith and Colles classification, - Salter - Harris classification, and - AO / OTA classification classify the anatomical abnormality according to any one of them.

[0018] In some examples, the medical image is an X - ray image.

[0019] Another aspect of the present invention provides a computer program product comprising instructions for enabling a processor to execute an embodiment of the above - described method.

[0020] Yet another aspect of the present invention provides a system for detecting anatomical abnormalities in 2D medical images. The system - acquires a 2D medical image, - detects a 2D contour of an anatomical feature, - acquires a 3D model representing the anatomical feature, - Predicting 2D contours of anatomical features based on a 3D model, where predicting 2D contours of anatomical features based on a 3D model means, - Estimating anatomical features and posture from 2D medical images. - Adjusting the 3D model based on the estimated posture, -Generating a 2D projection from a posture-adjusted 3D model. - Predicting 2D contours of anatomical features from 2D projections. Includes, - Detect anatomical abnormalities based on detected and predicted 2D contours. It has a processor configured in such a way.

[0021] In one example, the system may further include an X-ray source and a detector configured to acquire X-ray images.

[0022] These and other aspects of the present invention will become apparent from and be explained with reference to the embodiments described below. [Brief explanation of the drawing]

[0023] [Figure 1] A flowchart using One Halogen is shown. [Figure 2A] A medical image according to an embodiment of the present invention is shown. [Figure 2B] This shows the detected contours of anatomical features in medical images according to an embodiment of the present invention. [Figure 2C] This shows the predicted contours of anatomical features in medical images according to one embodiment of the present invention. [Figure 3] The posture and field of view parameters of an imaged object are shown using one embodiment of the present invention. [Figure 4] This shows a 3D model of the ankle joint according to an embodiment of the present invention. [Figure 5A] This shows the classification of ankle fractures according to embodiments of the present invention. [Figure 5B] This shows the classification of femoral neck fractures according to embodiments of the present invention. [Figure 5C] This describes the classification of fractures passing through the epiphysis or growth plate of a bone according to embodiments of the present invention. [Figure 6] This shows a processor circuit according to an embodiment of the present invention. [Modes for carrying out the invention]

[0024] The present invention is described herein with reference to the drawings. The description and specific examples provided illustrate exemplary embodiments, but should be understood to be for illustrative purposes only and not intended to limit the scope of the invention. Furthermore, the drawings are schematic diagrams and are not drawn to scale. Also, the same reference numerals are used throughout the drawings to indicate the same or similar parts.

[0025] The present invention provides a computer-aided method for analyzing and evaluating 2D (two-dimensional) medical images for the purpose of identifying anatomical abnormalities.

[0026] Figure 1 shows an exemplary flowchart of an embodiment of Method 100 according to the present invention, which includes the following steps.

[0027] In step 110, a 2D medical image 210 is acquired. The 2D medical image may be any type of medical image, including, but not limited to, X-ray images, CT (computed tomography) images, ultrasound images, MR (magnetic resonance) images, or any other type of medical image.

[0028] An exemplary medical image 210 envisioned in the present invention is a 2D X-ray image of the ankle joint 211 shown in Figure 2A.

[0029] In step 120, 2D contours of anatomical features are detected. This may include, for example, segmentation of 2D medical images or an end-to-end machine learning approach.

[0030] In general, segmentation can be understood as the task of classifying an image pixel by pixel into labels. Alternatively, the classification may be by larger image regions as a whole, such as contours or surfaces. The labels represent the semantics of each pixel or image region. In this way, in the classification operation, for example, bone pixels, i.e., pixels with the label "bone," can be identified. In this case, the pixels are thought to represent bone tissue, and so for other tissue types. Thus, the footprints of organs in an image can be identified by segmentation. In alternative embodiments, segmentation may be performed using a neural network or convolutional neural network, such as those disclosed in WO 2022 / 084074. Algorithms including prior shape knowledge, such as those disclosed therein, are also assumed by the present invention.

[0031] From image segmentation, a segmented image is typically obtained, where each pixel and / or region has a specific label assigned to it. This label may be binary "bone / non-bone" or multilevel "bone / ligament / non-bone / ...". According to one embodiment of the present invention, a portion of the medical image corresponding to an anatomical feature of interest, e.g., "bone" or more specifically "tibia / fibula / calcaneus / talus", may be selected to generate a 2D contour 212 as shown in Figure 2B.

[0032] In step 130, a 3D (three-dimensional) model 330 representing the imaged object 211 or identified anatomical feature 213 is acquired. The model is - Medical image 210 obtained in step 110, - Anatomical features identified in step 120, and - User input A 3D model can be selected from a list based on at least one of the following criteria.

[0033] For example, the image 210 acquired in step 110 may be annotated to indicate that the image is of the ankle joint. In another example, an identified anatomical feature or combination of identified anatomical features may indicate that the image includes at least the ankle joint. For example, a DICOM tag may be associated with the imaged anatomical feature, which can be used to identify the 3D model. In yet another example, a user of a computing system performing method 100 may select a 3D model from a list of models. The user selection may be from the entirety of all 3D models, or from a reduced set of articulated 3D models, which may be pre-selected based on 2D medical images and / or identified anatomical features within the 2D medical images. In yet another example, the 3D model may be selected based on contextual information in the acquisition location, e.g., a hospital radiology department, e.g., user diagnostic request data and / or standard radiological operating procedures for the requested examination procedure. For example, before image acquisition, the user or operator may indicate a desire to take an ankle joint X-ray image, which may determine the relevant 3D model or create a pre-selection of the relevant 3D model. The 3D model may be articulated so as to be able to take into account different postures of the imaged object; for example, the 3D model may be able to reproduce a specific flexion of the ankle joint.

[0034] A representative 3D model may be a computer-aided design (CAD) model created with any CAD software, or a reference model extracted from 3D imaging using known techniques. For example, a series of MR images of the ankle joint of a healthy subject may be acquired at various flexion angles. One image may then be segmented manually or automatically according to known techniques. The segmented images may then be used to create a surface model, for example, using a triangular mesh. For each triangle in the mesh, features for model adaptation may then be trained, for example, using image intensity, and the resulting model may be used to depict the corresponding bone surface in other MR images. This procedure may generate a 3D model for each MR image. The posture in each image and / or the corresponding 3D model may be further determined, for example, by measuring the angle between the principal axes of the tibia and the calcaneus for flexion. Smooth interpolation of the surface model between the changing postures recorded in the MR images may allow mesh constellations to be estimated for any posture between them.

[0035] A 3D model 430 envisioned by the present invention is shown in Figure 4. The 3D model may be articulated and may be modified to fit any posture encountered in 2D medical images. For example, throughout the image sequence in Figure 3, the calcaneus is shown to rotate from 80 degrees to 135 degrees of flexion relative to the tibia.

[0036] In step 140, the 2D contours of anatomical features in the medical image are predicted from the 3D model. The contours are then determined in the following steps, namely: -D Step 142 estimates the field of view parameters and / or posture of the captured object from the medical image, -2 Step 144 of modifying or adjusting the 3D model according to estimated field of view parameters and / or pose, - Step 146 of creating a 2D projection from a 3D model with adjusted field of view and / or posture, wherein the projection represents a 2D medical image 210, -Step 148 generates 2D contours 214 of anatomical features shown in the 2D projection of the 3D model, It is predicted according to this.

[0037] The field of view parameters and posture referenced in step 142 are, for example, as shown in Figure 3 for the ankle joint, - Flexion angle δ between the tibia and calcaneus; -Line of sight directions, including azimuthal rotation and polar rotation; and, -Detector position, translation, rotation, and scaling in the case of X-ray images Any one of these is acceptable.

[0038] This list should not be interpreted as exhaustive, and other field of view and / or posture parameters are also assumed by this invention.

[0039] Pose and / or field of view parameters may be estimated in step 142 according to S. Kronke et al. 2022 "CNN-based pose estimation for assessing quality of ankle-joint X-ray images".

[0040] For example, a mesh 3D model of an ankle joint may be sampled at random flexion angles and line-of-sight directions, and the projection may be generated by identifying triangles of the mesh in the 3D model traversed by an X-ray beam approximately tangentially to the selected line-of-sight direction, projecting their edges onto a virtual detector, semantically grouping them, and discretizing them on the detector pixel matrix. The resulting set of discretized contours in the projection forms the input for a pose estimation network. Such a pose estimation network may be a standard regression network including a ReLU (Rectified Linear Unit) activation function with alternating convolutional blocks of stride 1 and stride 2 and fully connected layers. The network, then trained according to the above principle, can be used to predict the pose of the detected 2D contour. For example, the captured 2D contour of the ankle joint is fed into a neural network, which, through its training, may provide the captured flexion angle and line-of-sight direction of the ankle joint.

[0041] Next, the posture and / or field of view parameters estimated in step 142 can be used as input to step 144 when modifying or adjusting the 3D model to reproduce the respective posture and / or field of view parameters. For example, the flexion angle of the 3D ankle model 430 may be changed as shown in Figure 4, and the angle between the tibia and calcaneus may be changed from 80 degrees to 135 degrees.

[0042] After adjusting the 3D model to the estimated pose, the 2D projection is determined in step 146 by, for example, identifying triangles of the mesh in the 3D model that are traversed by the X-ray beam approximately tangentially to the selected line of sight direction, projecting their edges onto a virtual detector, semantically grouping them, discretizing them onto the detector pixel matrix, thereby directly providing the predicted 2D contour in step 148. Figure 2C shows an exemplary 2D contour 214 predicted from the 3D model of the ankle joint in steps 140, 142 to 148.

[0043] In an optional step, the difference between a 2D contour generated based on a 2D medical image and a 2D contour generated based on an articulated 3D model may be determined. This difference may be calculated, for example, based on point-by-point subtraction or pixel-by-pixel subtraction. In some implementations, a difference map may be created, which may also be visualized on a user interface (not shown).

[0044] In step 150, anatomical abnormalities are detected based on the 2D contours detected in step 120 and predicted in step 140. Anatomical abnormalities may also be detected, for example, by visual inspection of a difference map shown on a user interface (not shown). In another exemplary embodiment, anatomical abnormalities may be detected automatically, for example, based on whether the difference or deviation between (i) 2D contours generated based on a 2D medical image and (ii) 2D contours generated based on an articulated 3D model exceeds a certain threshold. In yet another exemplary embodiment, a machine learning algorithm may be used to detect anatomical abnormalities. For example, 2D medical images may be annotated by a user or expert to identify anatomical abnormalities. These annotated images may be processed according to steps 120 and 140 to detect and predict 2D contours, where the detected 2D contours, predicted 2D contours, and / or the difference between the detected 2D contours and the predicted 2D contours may be added to the annotated images for use in the training process.

[0045] The identification of anatomical abnormalities may be binary, presence or absence of abnormality, or non-binary, in which case the probability of the abnormality being present may be provided. In an advantageous embodiment, anatomical abnormalities may also be classified in optional step 160 based on size, location, or any other form of classification to indicate a level of severity. That is, anatomical abnormalities may be classified into multiple categories. As will be discussed in more detail below, Figure 5A shows, for example, a classification of ankle fractures, and as will be discussed in more detail below, Figure 5B shows, for example, a classification of femoral neck fractures. According to one embodiment of the present disclosure, the classification of anatomical abnormalities is performed using a trained machine learning algorithm. For example, a classification network may be trained using annotated data such as those in Figures 5A to 5C, and using detected 2D contours 212 and predicted 2D contours 214 as input.

[0046] Annotated medical images are provided for training a classification machine learning model. For example, an ankle X-ray image has annotations indicating whether a fracture is present and the classification of the fracture according to the Weber classification, for example, Figure 5A. Annotation may be performed by a professional such as a clinician, radiologist, researcher, or any other qualified person. As an example, the annotated image may be an ankle X-ray image 210 with the relevant label indicating no fracture. Another example may be an ankle X-ray image with a fracture and a corresponding label identifying the classification of the fracture, for example, whether the fracture is a type A, B, or C fracture according to the Weber classification.

[0047] Similar to the images analyzed by Method 100, the annotated images used for training are also, -Detect 2D contours of anatomical features, - Predict 2D contours of anatomical features from healthy criteria, for example, from the 3D model obtained in step 130. To do so, follow steps 120 and 140 of Method 100.

[0048] The detected and predicted 2D contours can then form the input to a machine learning model, and the annotations on the training images can form the output. As is typical, training is performed using a subset of the training images and their corresponding annotations, and the remainder of the training images and their corresponding annotations are used to evaluate the trained model. Training is stopped when a predetermined criterion is reached, such as to ensure that the model is well-trained but not over-trained. In some implementations, noise may also be added to the training data, either manually or automatically, to avoid overfitting.

[0049] Although Method 100 is presented in a specific order, it will be understood by those skilled in the art that the order of the steps may be changed, additional steps may be added in between, and / or steps may be omitted.

[0050] Furthermore, it is understood that the individual steps in Method 100 may be performed in real time, for example, during or after the examination procedure. For example, 2D medical images may be acquired in real time and analyzed according to Method 100 during the image acquisition procedure, or 2D medical images may be acquired first during the acquisition step and analyzed only at the request of the operator and / or specialist. Thus, Method 100 may be performed within the image acquisition system or outside the system in a separate computing unit. In another embodiment, Method 100 may be performed in the cloud.

[0051] Predefined classifications envisioned by the present invention are disclosed in Figures 5A to 5C. The classifications in these figures are merely illustrative and do not limit the ideas of the present invention.

[0052] Figure 5A shows the Weber classification of the ankle joint. Figure 5A shows three different types of fractures, Type A 511, Type B 512, and Type C 513, which can be described as follows according to Wikipedia (https: / / en.wikipedia.org / wiki / Danis%E2%80%93Weber_classification). -Type A 511- Distal fibula fracture of the ligamentous joint (connection between the distal ends of the tibia and fibula). Typical features include: - Below the level of the ankle joint - Undamaged tibiofibular joint - Undamaged deltoid ligament - The medial malleolus is sometimes fractured. - Usually stable: Occasionally, however, open reduction and internal fixation (ORIF) may be necessary, especially if the medial malleolus is fractured. That is the case. -Type B 512- Fibular fracture at the level of the ligamentous joint. Typical features: - At the level of the ankle joint, extending above and laterally to the fibula -Tibiofibular joint that is intact or partially torn but does not widen the distal tibiofibular joint. - There is a possibility of a fracture of the medial malleolus or a rupture of the deltoid ligament. - Variable stability. -Type C 513- Proximal fibula fracture of the ligamentous joint. Typical features: - Above the level of the ankle joint - Tibiofibular ligament joint destroyed by dilation of the distal tibiofibular joint - Presence of medial malleolus fracture or deltoid ligament injury - Unstable: Requires ORIF.

[0053] Type B512 and Type C513 fractures indicate the extent of damage to the ligamentous joint itself (which may not be directly visualized on X-ray). They are inherently unstable and more likely to require surgical repair to achieve a good outcome. Type A fractures are usually stable and can be treated with simple measures such as plaster casts. Therefore, accurate identification of ankle fractures according to the Weber classification can improve the outcome and / or recovery of the subjects imaged.

[0054] Figure 5B shows the Pauwels classification of femoral neck fractures. These fractures are characterized by the fracture angle. Increasing angles result in more unstable fractures and increased shear stress at the fracture site. This shear leads to a higher rate of non-union. According to Wikipedia (https: / / en.wikipedia.org / wiki / Pauwel%27s_angle), the Pauwels classification divides fractures into Type I 521, Type II 522, and Type III 523 fractures, as follows: -Type I 521 -Fracture angle less than 30 degrees -Type II 522-30 to 50 degree fracture angle -Type III 523 -Fracture angle of 50° or greater.

[0055] Figure 5C shows the Salter-Harris classification of fractures involving the bone epiphysis or growth plate, specifically the zone of calcification. Thus, it is a form of pediatric fracture, occurring in 15% of long bone fractures in children. There are nine types of Salter-Harris fractures; types I 531 to V 535 (shown in Figure 5C), as described by Robert B Salter and W Robert Harris in 1963, and rarer types VI to IX (not shown or discussed here) that were added later. According to Wikipedia (https: / / en.wikipedia.org / wiki / Salter%E2%80%93Harris_fracture), the first five types can be described as follows: -Type I 531- Transverse fracture passing through the growth plate (also called the "epiphysis"); -Type II 532- Fractures that preserve the epiphysis, passing through the growth plate and metaphysis: Healing takes approximately 12 to 90 weeks or more in the spine; -Type III 533-Fractures passing through the growth plate and epiphysis, preserving the metaphysis; -Type IV 534- A fracture that passes through all three elements: bone, growth plate, metaphysis, and epiphysis; -Type V 535- Compression fracture of the growth plate (resulting in a perceived reduction in the space between the epiphysis and metaphysis on X-ray).

[0056] The predefined classifications in Figures 5A to 5C are provided as criteria for illustrating possible classifications according to embodiments of the present invention. In particular, using predefined classes allows operators to see not only whether a fracture may have occurred, but also what type of fracture may have occurred, and thus can improve subsequent steps to improve the outcome and / or recovery of the imaged subject.

[0057] Another predefined classification may be the AO / OTA (Arbeitsgemeinschaft fur Osteosynthesefragen / Orthopaedic Trauma Association) classification, which defines fractures according to their anatomical location, fracture type, and fracture subgroup. - Anatomical locations are described according to the AO / OTA classification, which is indicated by a two-digit number. The first digit indicates the major anatomical division (e.g., upper arm (humerus), lower arm, femur, lower leg (tibia / fibula), etc.), and the second digit indicates the specific part of the bone involved (e.g., proximal, central, distal).

[0058] - Fracture types are coded as type A, B, or C. -Type A indicates a simple fracture; -Type B indicates a wedge fracture with more than two fragment buses until there is some contact between the main fragments. - Type C indicates a complex or comminuted fracture with potentially multiple isolated fragments.

[0059] A fracture subgroup indicated by a single digit provides additional details regarding the nature of the fracture, such as the fracture pattern or composition (e.g., the fracture is an oblique fracture, transverse fracture, spiral fracture, etc.).

[0060] While certain classifications relating to bone and / or ligament fractures have been described above, other types of classifications, including but not limited to cancer types and / or stages, are also envisioned by the present invention. For example, a segmented tumor may be automatically classified into stage I, stage II, stage III, or stage IV tumors according to its size and / or location.

[0061] Method 100 may further include an optional step 170 in which treatment advice is provided based on the identified anatomical abnormalities and / or the classification results of the identified anatomical abnormalities. For example, treatment advice may be immediate surgery required for ankle fractures of type C Weber classification, while rest and / or cast placement may be advised for hairy and / or stress fractures. The advice may also be based on annotated data, for example, classification images may also be annotated to provide treatment advice according to the classification described above.

[0062] Figure 6 is a schematic diagram of a processor circuit 600 according to an embodiment of the present disclosure. As shown, the processor circuit 600 may include a processor 606, a memory 603, and a communication module 608. These elements may communicate with each other directly or indirectly, for example, via one or more buses.

[0063] The processor 606 envisioned by this disclosure may include a central processing unit (CPU), a graphics processing unit (GPU), a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a controller, a field-programmable gate array (FPGA) device, another hardware device, a firmware device, or any combination thereof configured to perform the operations described herein. The processor 606 may also be implemented as a combination of computing devices, for example, a DSP and a microprocessor, multiple microprocessors, a combination of one or other microprocessors working with a DSP core, or any other such configuration. The processor 606 may also implement various deep learning networks, which may include hardware or software implementations. The processor 606 may further include a preprocessor in either a hardware or software implementation.

[0064] The memory 603 envisioned by this disclosure may be any suitable storage device, such as a cache memory (e.g., the cache memory of the processor 606), random access memory (RAM), magnetoresistive RAM (MRAM), read-only memory (ROM), field-programmable gate array read-only memory (PROM), erasable field-programmable gate array read-only memory (EPROM), electrically erasable field-programmable gate array read-only memory (EEPROM), flash memory, solid-state memory devices, hard disk drives, other forms of volatile and non-volatile memory, or combinations of different types of memory. The memory may be distributed across multiple memory devices and / or located remotely from the processor circuitry. In one embodiment, the memory 603 may store instructions 605. Instructions 605 may include instructions that, when executed by the processor 606, cause the processor 606 to perform the operations described herein.

[0065] Instruction 605 may also be referred to as code. The terms “instruction” and “code” should be interpreted broadly to include any type of computer-readable statement. For example, the terms “instruction” and “code” may refer to one or more programs, routines, subroutines, functions, procedures, etc., and “instruction” and “code” may include a single computer-readable statement or many computer-readable statements. Instruction 205 may take the form of an executable computer program or script. For example, routines, subroutines, and / or functions may be defined in programming languages, including but not limited to C, C++, C#, Pascal, BASIC, API calls, HTML, XHTML, XML, ASP scripts, JavaScript, FORTRAN, COBOL, Perl, Java, ADA, NET, etc. Instructions may also include instructions for machine learning and / or deep learning.

[0066] The communication module 608 may include any electronic and / or logic circuits to facilitate direct or indirect communication of data between the processor circuit 600 and, for example, an external display (not shown) and / or an imaging device or system such as an X-ray imaging system. In this regard, the communication module 608 may be an input / output (I / O) device. Communication may be performed via any suitable means. For example, the communication means may be a wired link such as a Universal Serial Bus (USB) link or an Ethernet® link. Alternatively, the communication means may be a wireless link such as an ultra-wideband (UWB) link, an IEEE 802.11 Wi-Fi link, or a Bluetooth link.

[0067] The embodiments described above are illustrative, not limiting, of the invention, and it should be noted that those skilled in the art can design many alternative embodiments without departing from the scope of the appended claims. No reference numeral placed in parentheses in a claim should be construed as limiting that claim. The term “having” does not exclude the existence of elements or steps other than those enumerated in the claim. The word “a” or “an” preceding an element does not exclude the existence of multiple such elements. It may be implemented by hardware having several distinct elements and / or by a appropriately programmed processor. In an apparatus claim enumerating several means, some of these means may be embodied by the same hardware item. Means described in different dependent claims may be used in combination for advantage.

Claims

1. A computer-based method for detecting anatomical abnormalities in 2D medical images using a processor, The processor performs the steps of acquiring the 2D medical image, The processor performs the steps of detecting the 2D contour of an anatomical feature, The processor takes the step of acquiring a 3D model representing the anatomical features, The step of the processor predicting the 2D contour of the anatomical features based on the 3D model, the step of predicting the 2D contour of the anatomical features based on the 3D model, A step of estimating the posture of the anatomical features from the 2D medical image, A step of adjusting the 3D model based on the estimated posture, The steps include generating a 2D projection from the adjusted 3D model, and A step of predicting the 2D contour of the anatomical feature from the 2D projection, Steps including, The processor performs the steps of detecting the anatomical abnormality based on the detected 2D contour and the predicted 2D contour, A method of having.

2. The step of detecting the 2D contour is, The step of segmenting the aforementioned 2D medical image, The method according to claim 1, including the method described in claim 1.

3. The processor performs the step of classifying the anatomical abnormalities. The method according to claim 1, further comprising:

4. The method according to claim 3, wherein the step of classifying the anatomical abnormalities is performed by a machine learning algorithm.

5. The aforementioned 3D model is The aforementioned 2D medical image, The aforementioned anatomical features, and / or User input, The method according to claim 1, obtained based on the present invention.

6. The aforementioned 3D model, Computer-aided design models, and A reference model constructed from 3D medical images of the anatomical features in at least one posture, The method according to claim 1, wherein the method is any one of the following.

7. The detected 2D contour and, The predicted 2D contour line and, A step to generate the difference between, The method according to claim 1, further comprising:

8. The aforementioned anatomical features, Bone tissue, and Ligamentous tissue, The method according to claim 1, wherein the method is any one of the following.

9. The aforementioned anatomical abnormalities Fractures, and Ligament rupture, The method according to claim 1, wherein the method is any one of the following.

10. The step of classifying the aforementioned anatomical abnormalities is: Weber classification, Pauwels classification, Smith and Colles classification, Salter-Harris classification, and AO / OTA classification, The method according to claim 3, wherein the anatomical abnormality is classified based on any one of the following.

11. The processor generates treatment advice based on the detected anatomical abnormality. The method according to claim 1, further comprising:

12. The method according to claim 1, wherein the medical image is an X-ray image.

13. A computer program having instructions for enabling a processor to perform the method according to any one of claims 1 to 12.

14. A system for detecting anatomical abnormalities in 2D medical images, comprising a processor configured to perform the method according to any one of claims 1 to 12.

15. The system according to claim 14, further comprising an X-ray source and an X-ray detector.

Images