Method for integrally analyzing spinal diseases, and electronic device for performing same

Abstract

A method by which an electronic device integrally analyzes spinal diseases, according to one embodiment of the present application, may comprise the steps of: receiving a spinal image; applying the spinal image to a trained vertebrae segment detection model so as to detect a plurality of vertebrae segments; measuring the height and the width of a plurality of vertebrae and the intervals between the plurality of vertebrae on the basis of the vertebrae segments so as to generate measurement indexes; applying the vertebra segments to one or more trained image disease analysis models so as to generate disease information about the vertebra; applying the measurement indexes and the disease information to a trained integrated analysis model so as to generate integrated disease information about the vertebra; and outputting the integrated disease information. According to one embodiment of the present invention, an X-ray spinal image is integrally analyzed to determine a spinal disease area and the type of the corresponding spinal disease.

Description

Method for integrated analysis of spinal diseases and electronic device for performing the same

[0001] The present application relates to a method for integrated analysis of spinal diseases and an electronic device for performing the same. Specifically, the present application relates to a method for analyzing spinal diseases by integrating a method for measuring the structure of the vertebrae and a method for analyzing spinal diseases visually using an artificial intelligence model, and to an electronic device for performing the same.

[0002] As artificial intelligence technology advances, the field of medical image analysis, which analyzes medical images to produce diagnostic indicators related to various diseases, is attracting attention.

[0003] Medical imaging has advantages and limitations depending on the method used to capture the images. For example, X-ray equipment offers the advantages of rapid imaging, low cost, and short analysis times, making it a primary diagnostic tool in most medical settings. When spinal disease is suspected, X-ray imaging is often the first-line test.

[0004] However, it is often difficult to detect diseases by visually examining two-dimensional cross-sectional images of X-ray images, and there is a problem that there is a large deviation in the diagnosis of diseases depending on the difference in reading experience and the condition of the reader.

[0005] Although CT or MRI scans are necessary for accurate diagnosis, they are expensive and time-consuming. Therefore, X-ray imaging is often used as the initial testing method. If no disease is suspected at this stage, the patient often does not proceed to the next stage (CT or MRI). Therefore, the diagnosis process using X-ray images is crucial.

[0006] Recent machine learning-based automated analysis methods rely on conventional methods, such as classifying images for disease or detecting diseased areas within the images. These simple methods have limited accuracy in identifying disease in standard X-ray images.

[0007] Accordingly, the problem to be solved by the present invention is to provide a method for comprehensively analyzing spinal diseases and an electronic device for performing the same.

[0008] The problems to be solved by the present invention are not limited to the problems described above, and problems not mentioned can be clearly understood by a person having ordinary skill in the technical field to which the present invention pertains from this specification and the attached drawings.

[0009] According to an embodiment of the present invention, an electronic device may include a method for integrating and analyzing a spinal disease, the method including the steps of: receiving a spinal image; detecting a plurality of spinal segments by applying the spinal image to a learned spinal segment detection model; measuring heights and widths of the plurality of vertebrae and gaps between the plurality of vertebrae based on the spinal segments to generate measurement indices; applying the spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal bone; applying the measurement indices and disease information to a learned integrated analysis model to generate integrated disease information of the spinal bone; and outputting the integrated disease information.

[0010] According to one embodiment of the present invention, a device for integrated spinal disease analysis includes a communication unit, a processor, and a memory that inputs a spinal image and outputs integrated spinal disease information, wherein the processor may include a spinal segment detection unit that detects a plurality of spinal segments by applying the spinal image to a learned spinal segment detection model, a spinal measurement unit that measures the height and width of a plurality of vertebrae and the gaps between the plurality of vertebrae based on the spinal segments to generate measurement indices, an imaging disease analysis unit that applies the spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal bone, and an integrated analysis unit that applies the measurement indices and the disease information to a learned integrated analysis model to generate integrated disease information of the spinal bone.

[0011] The solutions to the problems of the present invention are not limited to the solutions described above, and solutions that are not mentioned can be clearly understood by a person having ordinary skill in the art to which the present invention pertains from this specification and the attached drawings.

[0012] According to one embodiment of the present invention, by comprehensively analyzing X-ray spinal images, the type of spinal disease and the area of ​​the spinal disease can be accurately determined.

[0013] FIG. 1 is a block diagram briefly showing the configuration of an electronic device according to one embodiment of the present application.

[0014] FIG. 2 is a drawing for explaining various operations of an electronic device according to one embodiment of the present application.

[0015] FIG. 3 is a drawing for explaining various operations of a video disease analysis unit according to one embodiment of the present application.

[0016] FIG. 4 is a drawing for explaining an artificial intelligence model of an electronic device according to one embodiment of the present application.

[0017] FIG. 5 is a flowchart illustrating a method for comprehensively analyzing spinal diseases according to one embodiment of the present application.

[0018] FIG. 6 is a drawing for explaining a method for detecting a spinal segment according to one embodiment of the present application.

[0019] FIG. 7 is a drawing for explaining a method for detecting a spinal segment according to one embodiment of the present application.

[0020] FIG. 8 is a diagram showing a detected spinal segment according to one embodiment of the present application.

[0021] FIG. 9 is a diagram illustrating a method for outputting integrated disease information according to one embodiment of the present application.

[0022] The above-described purposes, features, and advantages of the present application will become more apparent through the following detailed description, taken in conjunction with the accompanying drawings. However, as the present application is susceptible to various modifications and various embodiments, specific embodiments will be illustrated in the drawings and described in detail below.

[0023] Throughout the specification, identical reference numbers, in principle, indicate identical components. Furthermore, components with identical functions within the scope of the same concept shown in the drawings of each embodiment are described using the same reference numbers, and redundant descriptions thereof will be omitted.

[0024] If a detailed description of a known function or configuration related to this application is deemed to unnecessarily obscure the gist of this application, such detailed description will be omitted. Furthermore, numbers (e.g., "first," "second," etc.) used throughout the description of this specification are merely identifiers used to distinguish one component from another.

[0025] In addition, the suffixes "module" and "part" for components used in the following examples are given or used interchangeably only for the convenience of writing the specification, and do not have distinct meanings or roles in themselves.

[0026] In the examples below, singular expressions include plural expressions unless the context clearly indicates otherwise.

[0027] In the examples below, terms such as “include” or “have” mean that a feature or component described in the specification is present, and do not preclude the possibility that one or more other features or components may be added.

[0028] For convenience of explanation, the sizes of components in the drawings may be exaggerated or reduced. For example, the sizes and thicknesses of each component shown in the drawings are arbitrarily shown for convenience of explanation, and the present invention is not necessarily limited to what is shown.

[0029] In some embodiments, where implementations are otherwise feasible, the order of specific processes may differ from the order described. For example, two processes described in succession may be performed substantially simultaneously, or in a reverse order from the order described.

[0030] In the following examples, when components are said to be connected, this includes not only cases where the components are directly connected, but also cases where components are interposed between the components and are indirectly connected.

[0031] For example, when it is said in this specification that components, etc. are electrically connected, it includes not only cases where the components, etc. are directly electrically connected, but also cases where components, etc. are interposed in between and are indirectly electrically connected.

[0032] The present invention is applicable to a wide range of fields. While this specification focuses on a method for analyzing spinal diseases, the present invention is not limited to any specific body part and can be utilized in a variety of medical analyses for analyzing bone diseases.

[0033] The AI ​​model according to the present invention may be a machine learning-based deep learning model. Furthermore, these models may be trained using acquired data or statistical data. Furthermore, the division of AI models in the present invention is merely intended to enhance understanding; in reality, there may be one or more models.

[0034]

[0035] Hereinafter, a method for integrating analysis of a spinal disease of the present application and an electronic device performing the same will be described with reference to FIGS. 1 and 6. In this specification, it is assumed and described that the spinal image is an X-ray image of vertebrae taken using an X-ray device. In addition, the spinal image may be part or all of a frontal spinal image, a lateral spinal image, or a posterior spinal image. However, this is merely an example for convenience of explanation and is not to be construed as limiting thereto. Therefore, the embodiments described below may be analogically applied to other types of medical images other than X-ray images of the spinal image.

[0036]

[0037] FIG. 1 is a block diagram briefly showing the configuration of an electronic device according to one embodiment of the present application. The electronic device (100) according to one embodiment of the present application receives a spinal image, applies the spinal image to a learned spinal segment detection model to detect a plurality of spinal segments, measures the height and width of the plurality of vertebrae and the gaps between the plurality of vertebrae based on the spinal segments to generate measurement indices, applies the spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal bone, applies the measurement indices and the disease information to a learned integrated analysis model to generate integrated disease information of the spinal bone, and displays and outputs the integrated disease information on the spinal image.

[0038]

[0039] Referring to FIG. 1, an electronic device (100) may include a communication unit (110), a processor (120), and a memory (130).

[0040]

[0041] The communication unit (110) may support the establishment of a direct (e.g., wired) communication channel or a wireless communication channel between the electronic device (100) and an external electronic device (e.g., a radiographic imaging (X-ray) photographing device), and the performance of communication through the established communication channel. The communication unit (110) may include one or more communication processors that operate independently from the processor (120) (e.g., an application processor) and support direct (e.g., wired) communication or wireless communication. According to one embodiment, the communication unit (110) may include a wireless communication module (e.g., a cellular communication module, a short-range wireless communication module, or a global navigation satellite system (GNSS) communication module) or a wired communication module (e.g., a local area network (LAN) communication module, or a power line communication module). Any of these communication modules can communicate with external electronic devices via a short-range communication network such as Bluetooth, WiFi (Wireless Fidelity) Direct, or IrDA (Infrared Data Association), or a long-range communication network such as a legacy cellular network, a 5G network, a next-generation communication network, the Internet, or a computer network (e.g., a LAN or WAN). These various types of communication modules can be integrated into a single component (e.g., a single chip) or implemented as multiple separate components (e.g., multiple chips).

[0042] The communication unit (110) can receive a spinal image from an external device. For example, the communication unit (110) can receive an X-ray image including a spinal bone photographed by an X-ray device.

[0043] The communication unit (110) can transmit the integrated disease information analyzed in the electronic device (100) to an external device. At this time, the integrated disease information may be information on the type of spinal disease and the area of ​​the spinal disease.

[0044]

[0045] The processor (120) may execute software to control at least one other component (e.g., hardware or software component) of the electronic device (100) connected to the processor (120) and perform various data processing or calculations. According to one embodiment, as at least a part of the data processing or calculation, the processor (120) may store commands or data received from other components (e.g., the communication unit (110)) in a volatile memory, process the commands or data stored in the volatile memory, and store the data in a non-volatile memory. According to one embodiment, the processor (120) may include a main processor (e.g., a central processing unit or an application processor) or an auxiliary processor (e.g., a graphics processing unit, a neural processing unit (NPU), an image signal processor, a sensor hub processor, or a communication processor) that can operate independently or together therewith.

[0046] According to one embodiment, the auxiliary processor (e.g., a neural network processing unit) may include a hardware structure specialized for processing artificial intelligence models. The artificial intelligence models may be generated through machine learning. Such learning may be performed, for example, within the electronic device (100) on which the artificial intelligence is performed, or may be performed through a separate server (e.g., server (108)). The learning algorithm may include, for example, supervised learning, unsupervised learning, semi-supervised learning, or reinforcement learning, but is not limited to the examples described above. The artificial intelligence model may include multiple artificial neural network layers. The artificial neural network may be one of a deep neural network (DNN), a convolutional neural network (CNN), a recurrent neural network (RNN), a restricted Boltzmann machine (RBM), a deep belief network (DBN), a bidirectional recurrent deep neural network (BRDNN), deep Q-networks, a Vision Transformer series model, or a combination of two or more of the above, but is not limited to the examples described above. In addition to, or as an alternative to, a hardware structure, the artificial intelligence model may include a software structure.

[0047] The processor (120) can control the overall operation of the electronic device (100). For example, the processor (120) can control an operation of receiving a spinal image, an operation of detecting a plurality of spinal segments by applying the spinal image to a learned spinal segment detection model, an operation of measuring the height and width of a plurality of vertebrae and the gaps between the plurality of vertebrae based on the spinal segments to generate measurement indices, an operation of applying the spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal bone, an operation of generating integrated disease information of the spinal bone by applying the measurement indices and disease information to a learned integrated analysis model, and an operation of outputting the integrated disease information.

[0048]

[0049] The memory (130) can store various data used by at least one component (e.g., the processor (120)) of the electronic device (101). The data may include, for example, input data or output data for software and commands related thereto. The memory (130) may include volatile memory or non-volatile memory. The memory (130) may be provided in a form built into the electronic device (100) or in a form that is detachable.

[0050]

[0051] FIG. 2 is a diagram illustrating various operations of an electronic device according to one embodiment of the present application. FIG. 2 will be described in more detail with reference to FIG. 3. FIG. 3 is a diagram illustrating various operations of an imaging disease analysis unit according to one embodiment of the present application.

[0052]

[0053] Referring to FIG. 2, the electronic device (100) may include a spinal image input unit (210), a spinal segment detection unit (220), a spinal measurement unit (230), an imaging disease analysis unit (240), an integrated analysis unit (250), and an integrated analysis information output unit (260).

[0054]

[0055] The spinal image input unit (210) can receive a spinal image from an external device. For example, the spinal image input unit (210) can receive an X-ray image including a spinal bone photographed by an X-ray photographing device.

[0056]

[0057] The spinal segment detection unit (220) can detect spinal segments by applying a spinal image to an artificial intelligence model (spinal segment detection model). Specifically, the spinal segment detection unit (220) can detect each bone constituting the spinal column (seven cervical vertebrae, twelve thoracic vertebrae, five lumbar vertebrae, sacrum, and coccyx) by applying a spinal image to the artificial intelligence model.

[0058] For example, the vertebra segment detection unit (220) can detect vertebra segments in a rectangular shape in a vertebra image using an artificial intelligence model (e.g., Bounding Box detection). Specifically, the vertebra segment detection unit (220) (or any external device) can train a Bounding Box detection model based on a spine image and a label assigned to the spine image as a spine segment region. At this time, the label can be automatically assigned to the spine image using any software or can be manually assigned to the spine image by any operator. More specifically, the vertebra segment detection unit (220) can receive a spine image and be trained to minimize the difference between the output value and the label associated with the vertebra segment region.

[0059] For another example, the vertebra segment detection unit (220) can detect the edges of each vertebra in the vertebra image using an artificial intelligence model (e.g., keypoint detection). Specifically, the vertebra segment detection unit (220) (or any external device) can train a keypoint detection model based on the spine image and the labels assigned to the spine image as vertebra edges. At this time, the labels can be automatically assigned to the spine image using any software or can be manually assigned to the spine image by any operator. More specifically, the vertebra segment detection unit (220) can receive the spine image and be trained to minimize the difference between the output value and the label associated with the vertebra segment region.

[0060] In addition, the spine segment detection unit (220) can detect spine segments in a rectangular shape in a spine image using a bounding box detection model, and can also detect the corners of the spine within the rectangle by applying the spine segments included in each rectangle to a keypoint detection model.

[0061] Specific examples of AI models include convolutional neural networks (CNNs), recurrent neural networks (RCNs), deep neural networks (DNNs), and generative adversarial networks (GANs). However, these are merely examples and should be interpreted as a comprehensive term encompassing the aforementioned AI models, various other types of neural networks, and combinations thereof. This does not necessarily mean that they are necessarily part of the deep learning field.

[0062]

[0063] The spine measurement unit (230) can measure the height and width of a plurality of vertebrae and the spacing between the plurality of vertebrae based on the vertebrae segments. The spine measurement unit (230) can automatically measure the height, width, and interosseous spacing of each bone based on the indicators of the edges of the detected vertebrae. The spine measurement unit (230) can determine a bone or interosseous space where the height, width, and interosseous spacing of each bone deviate from the normal standard. At this time, information on the normal standard for the height, width, and interosseous spacing of each of the plurality of bones corresponding to the vertebrae may be stored in advance.

[0064] The spine measurement unit (230) can extract the characteristics of a suspected disease by comparing the bone structure measurement data with the bone structure measurement data having a specific disease. For example, in the case of a degenerative disease, the space between the vertebrae becomes narrow, so if the measurement result determines that the space between the bones is narrow, the spine measurement unit (230) can determine that a degenerative disease is suspected in the corresponding area. For another example, in the case of a fracture, a change in bone height occurs, so if the measurement result determines that the bone height is different, the spine measurement unit (230) can determine that a fracture is suspected in the corresponding area. At this time, the bone structure measurement data having a specific disease may be stored in advance.

[0065] The spine measurement unit (230) can automatically measure the degree of inclination of a vertebra based on the vertebra segments. For example, the spine measurement unit (230) can analyze changes in the position of a vertebra and detect a vertebra suspected of having a disease by analyzing the positional abnormalities of a specific bone.

[0066]

[0067] The imaging disease analysis unit (240) can generate disease information of the spine by applying the vertebrae segments to one or more learned imaging disease analysis models. The imaging disease analysis unit (240) can detect a first disease region in a rectangular shape by applying the vertebrae segments to one or more imaging analysis models. In addition, the imaging disease analysis unit (240) can determine a type of disease corresponding to the spine by applying the vertebrae segments to one or more imaging analysis models. In addition, the imaging disease analysis unit (240) can segment the disease region by applying the vertebrae segments to one or more imaging analysis models to detect a second disease region. The imaging disease analysis unit (240) can generate disease information of the spine based on at least one of the first disease region, the type of disease, and the second disease region.

[0068] Referring to FIG. 3, the imaging disease analysis unit (240) will be described in detail. Referring to FIG. 3, the imaging disease analysis unit (240) may include a preprocessing unit (241), a first disease area determination unit (243), a disease type determination unit (245), and a second disease area determination unit (247).

[0069]

[0070] The preprocessing unit (241) may perform preprocessing of spinal images for imaging disease analysis, which will be described later. For example, the preprocessing unit (241) may be implemented to perform an operation to correct the intensity of spinal images. Alternatively, the preprocessing unit (241) may perform an operation to remove noise from spinal images, normalize them, or align them.

[0071]

[0072] The first disease region determination unit (243) can detect a diseased region among the vertebrae segments included in the spinal image in a rectangular shape using an artificial intelligence model (e.g., a detection model). Specifically, the first disease region determination unit (243) (or any external device) can train a detection model based on the vertebrae segment images and labels that assign disease regions to the diseased areas. At this time, the labels can be automatically assigned to the spinal image using any software or manually assigned to the spinal image by any operator. More specifically, the first disease region determination unit (243) can receive the vertebrae segment images and be trained to minimize the difference between the output values ​​and the labels related to the disease regions among the vertebrae segments.

[0073] According to an embodiment, the first disease region determination unit (243) may detect the first disease region in a multi-region manner. Specifically, the first disease region determination unit (243) may detect multiple images from a spinal image, determine a disease region from each image, and determine overlapping disease regions. For example, the first disease region determination unit (243) may extract an image of the entire thoracic region, an image including multiple thoracic vertebrae, and an image including one thoracic vertebra from a spinal image based on spinal bone segments. Then, the first disease region determination unit (243) may determine a disease region from each of the extracted images. Then, the first disease region determination unit (243) may detect the first disease region by comprehensively analyzing the disease regions determined from each of the images. Specifically, by mapping probability values ​​for regions detected as diseased in the entire image of the thoracic region, regions detected as diseased in images containing multiple thoracic vertebrae, and regions detected as diseased in images containing one thoracic vertebrae, the probability of suspecting disease in regions with high combined probability values ​​can be increased. That is, the first disease region determination unit (243) can detect regions overlapping with disease regions in all of the multiple images as the first disease region with high probability.

[0074]

[0075] The disease type determination unit (245) can determine the type of disease corresponding to the vertebrae using an artificial intelligence model (e.g., a classification model). For example, the disease type determination unit (245) can detect the presence or absence of specific diseases such as infection, degenerative change, metastasis, and fracture. Specifically, the disease type determination unit (245) (or any external device) can train a classification model based on a specific disease image and a text indicating a specific disease with a specific label. More specifically, the disease type determination unit (245) can receive images of vertebrae segments and be trained to minimize the difference between the output value and the label related to the disease region among the vertebrae segments.

[0076] According to an embodiment, the disease type determination unit (245) may determine the disease type in a multi-region manner. Specifically, the disease type determination unit (245) may detect multiple images from a spinal image, determine the disease type from each image, and determine overlapping disease types. For example, the disease type determination unit (245) may extract an image of the entire thoracic region, an image including multiple thoracic vertebrae, and an image including one thoracic vertebra from a spinal image based on vertebrae segments. Then, the disease type determination unit (245) may determine the disease type corresponding to each of the extracted images. Then, the disease type determination unit (245) may determine the disease type by comprehensively analyzing the disease types determined from each of the images. Specifically, the probability values ​​for the disease types determined from the entire thoracic region image, the disease types determined from the images including multiple thoracic vertebrae, and the disease types determined from the images including one thoracic vertebrae may be mapped, thereby increasing the probability of the disease type having a high combined probability value. That is, the disease type determination unit (245) can determine the disease type determined from all of the multiple images as the disease type with a high probability.

[0077]

[0078] The second disease area determination unit (247) can segment and detect diseased areas among the spinal bone segments included in the spinal image using an artificial intelligence model (e.g., segmentation model). For example, the second disease area determination unit (247) can detect diseased areas by performing segmentation on areas suspected of having a specific disease within the spinal image.

[0079] Specifically, the second disease region determination unit (247) (or any external device) can train a segmentation model based on the images of the spinal segments and the labels assigned to the diseased areas. At this time, the labels can be automatically assigned to the spinal images using any software or manually assigned to the spinal images by any operator. More specifically, the second disease region determination unit (247) can receive the images of the spinal segments and be trained to minimize the difference between the output values ​​and the labels related to the diseased areas among the spinal segments.

[0080] According to an embodiment, the second disease region determination unit (247) may detect the second disease region in a multi-region manner. Specifically, the second disease region determination unit (247) may detect multiple images from a spinal image, determine a disease region from each image, and determine an overlapping disease region. For example, the second disease region determination unit (247) may extract an image of the entire thoracic region, an image including multiple thoracic vertebrae, and an image including one thoracic vertebra from a spinal image based on vertebra segments. Then, the second disease region determination unit (247) may determine a disease region from each of the extracted images. Then, the second disease region determination unit (243) may detect the first disease region by comprehensively analyzing the disease regions determined from each of the images. Specifically, by mapping probability values ​​for regions detected as diseased in the entire image of the thoracic region, regions detected as diseased in images containing multiple thoracic vertebrae, and regions detected as diseased in images containing one thoracic vertebrae, the probability of suspecting disease in regions with high combined probability values ​​can be increased. That is, the second disease region determination unit (247) can detect regions overlapping with disease regions in all of the multiple images as the second disease region with high probability.

[0081]

[0082] Referring back to FIG. 2, the integrated analysis unit (250) can generate integrated disease information of the spine by applying measurement indicators and disease information to the learned integrated analysis model. Specifically, the integrated analysis unit (250) can determine an integrated disease type and an integrated disease area based on the vertebrae or the gap between vertebrae that deviates from the normal standard determined by the spine measurement unit (230) and the disease area and disease type determined by the imaging disease analysis unit (240).

[0083] In addition, the integrated analysis unit (250) can integrate and analyze the analysis results for images from different directions. For example, the analysis results for the frontal image and the side image can be integrated to determine the presence, type, and location of a disease. Specifically, the integrated analysis unit (250) can determine whether a disease is present in the frontal image by considering the probability of disease suspicion in the area of ​​the side image corresponding to the area suspected of having a disease in the frontal image.

[0084] The integrated analysis information output unit (260) can output integrated disease information. For example, the integrated analysis information output unit (260) can display the integrated disease information on a spinal image and output it, or transmit it to an external device.

[0085] Additionally, the integrated analysis information output unit (260) can display integrated disease information on images of different orientations simultaneously or sequentially. For example, if a disease is determined to be present in the lumbar vertebrae 3, the integrated analysis information output unit (260) can map and display frontal and side images showing disease information on the lumbar vertebrae 3.

[0086]

[0087] FIG. 4 is a diagram illustrating an artificial intelligence model of an electronic device according to an embodiment of the present application. Referring to FIG. 4, the electronic device (100) can detect spinal segments using a segment detection model. In addition, the electronic device (100) can generate spinal disease information using an imaging disease analysis model. In addition, the electronic device (100) can generate integrated spinal disease information using an integrated analysis model. In the present invention, the segment detection model, the imaging disease analysis model, and the integrated analysis model are divided only to enhance understanding of the invention, and each model may actually be one or more models.

[0088]

[0089] FIG. 5 is a flowchart illustrating a method for comprehensively analyzing spinal diseases according to one embodiment of the present application. The operations in FIG. 5 are not limited in order, and additional operations may be performed between two adjacent operations. Furthermore, at least some of the operations in FIG. 5 may be omitted. In the present invention, the expression that the electronic device (100) performs a specific operation may mean that the processor (120) of the electronic device (100) performs the specific operation, or that the processor (120) controls other hardware to perform the specific operation.

[0090] FIG. 5 will be described in more detail with reference to FIGS. 6 and 7. FIG. 6 is a diagram illustrating a detected vertebral segment according to one embodiment of the present application. FIG. 7 is a diagram illustrating a method for outputting integrated disease information according to one embodiment of the present application.

[0091]

[0092] Referring to FIG. 5, the electronic device (100) can receive an image of the spine (S1000). For example, the electronic device (100) can receive an X-ray image including a spine photographed by an X-ray photographing device. The X-ray image including a spine may be a frontal X-ray image of the spine, a lateral X-ray image of the spine, or a posterior X-ray image of the spine.

[0093]

[0094] The electronic device (100) can detect multiple vertebrae segments by applying a spinal image to a learned vertebrae segment detection model (S2000). Specifically, the electronic device (100) can detect each bone constituting the spine (seven cervical vertebrae, twelve thoracic vertebrae, five lumbar vertebrae, sacrum, and coccyx) by applying the spinal image to the artificial intelligence model. For example, the electronic device (100) can detect the bones corresponding to each of cervical 1, cervical 2, .., thoracic 1, thoracic 2, .., lumbar 1, lumbar 2, .., sacrum, and coccyx in a rectangular shape or by using the four corners of each bone as landmarks. For example, referring to FIGS. 6(a) and 6(b), the electronic device (100) can detect each bone constituting the spine by using the four corners of each bone included in the spinal image as landmarks. Alternatively, the electronic device (100) may detect the edges of the bones within the rectangular area after detecting each bone in a rectangular shape. For example, as shown in FIG. 7, the electronic device (100) may detect the edges of the bones within the rectangular area after detecting each bone included in the spinal image in a rectangular shape. According to various embodiments, as shown in FIG. 8, the electronic device (100) may apply the spinal image to a spinal bone segment detection model to detect a plurality of spinal bone segments.

[0095]

[0096] The electronic device (100) can generate measurement indices by measuring the height and width of a plurality of vertebrae and the spacing between the plurality of vertebrae based on the vertebrae segments (S3000). Specifically, the electronic device (100) can automatically measure the height, width, inter-bone spacing, angle, etc. of each bone based on the edge indices of the vertebrae detected in the S2000 operation. The electronic device (100) can determine a bone or inter-bone where the height, width, inter-bone spacing, angle, etc. of each bone deviates from the normal standard. At this time, information on the normal standard for the height, width, inter-bone spacing, angle, etc. of each of the plurality of bones corresponding to the vertebrae may be stored in advance.

[0097]

[0098] The electronic device (100) can generate disease information of the spine by applying the spinal segments to one or more learned imaging disease analysis models (S4000). The electronic device (100) can detect a first disease region in a rectangular shape by applying the spinal segments to a detection model. In addition, the electronic device (100) can determine the type of disease corresponding to the spine by applying the spinal segments to a classification model. In addition, the electronic device (100) can detect a second disease region by segmenting the disease region by applying the spinal segments to a segmentation model. The imaging disease analysis unit (240) can generate disease information indicating the type and disease region of the spine based on at least one of the first disease region, the type of disease, and the second disease region.

[0099]

[0100] The electronic device (100) can generate integrated disease information of the spine by applying measurement indicators and disease information to a learned integrated analysis model (S5000). Specifically, the electronic device (100) can determine an integrated disease type and integrated disease area based on the vertebrae or the gap between vertebrae that deviates from the normal standard determined in operation S3000 and the disease area and disease type determined in operation S4000.

[0101]

[0102] The electronic device (100) can output integrated disease information (S6000). For example, the electronic device (100) can display and output the integrated disease information on a spinal image or transmit it to an external device. For example, as shown in FIG. 9, the electronic device (100) can display and output a disease area on a spinal image. Alternatively, the electronic device (100) can display and output the same disease area together in spinal image images captured from different angles, as shown in FIGS. 7(a) and 7(b).

[0103]

[0104] According to a method for integrated analysis of spinal disease according to an embodiment of the present invention and an electronic device for performing the same, there is an advantage in that the type of spinal disease and the area of ​​the spinal disease can be determined by integrated analysis of an X-ray spinal image.

[0105]

[0106] The features, structures, effects, etc. described in the embodiments above are included in at least one embodiment of the present invention, and are not necessarily limited to just one embodiment. Furthermore, the features, structures, effects, etc. exemplified in each embodiment can be combined or modified in other embodiments by those skilled in the art to which the embodiments pertain. Therefore, the contents related to such combinations and modifications should be construed as falling within the scope of the present invention.

[0107] In addition, although the above description focuses on the embodiments, these are merely examples and do not limit the present invention. Those skilled in the art to which the present invention pertains will appreciate that various modifications and applications not exemplified above are possible without departing from the essential characteristics of the present embodiments. In other words, each component specifically shown in the embodiments can be modified and implemented. In addition, differences related to such modifications and applications should be interpreted as being included within the scope of the present invention defined in the appended claims.

Claims

1. In a method for integrating analysis of spinal diseases by an electronic device, Step of receiving spinal images; A step of applying the above spinal image to a learned spinal segment detection model to detect multiple spinal segments; A step of generating measurement indices by measuring the height and width of a plurality of vertebrae and the spacing between the plurality of vertebrae based on the above spinal segments; A step of applying the above spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal bone; A step of applying the above measurement indicators and disease information to a learned integrated analysis model to generate integrated disease information of the spine; and A method for integrated spinal disease analysis, comprising: a step of outputting the integrated disease information.

2. In paragraph 1, The step of generating disease information of the spine by applying the above spinal segments to one or more learned imaging disease analysis models is: A step of applying the above spinal segments to a learned disease detection model to determine a first disease area; A step of applying the above spinal segments to a learned disease classification model to determine the type of disease corresponding to the spinal segment; A step of determining a second disease area by applying the above spinal segments to a learned area segmentation model; and A step of generating disease information of the spine based on the first disease area, the disease type, and the second disease area; Integrated analysis method for spinal diseases.

3. In paragraph 2, The step of applying the above spinal segments to the learned disease detection model to determine the first disease area in the spinal image is as follows: A step of extracting multiple regions of interest from the above spinal image; A step of determining a disease area in each of the above multiple areas of interest: and A step of determining an area where the disease areas determined in each of the plurality of areas of interest overlap as a first disease area; Integrated analysis method for spinal diseases.

4. In paragraph 3, The step of extracting multiple regions of interest from the above spinal image is: A step of extracting a lumbar image from the above spinal image; A step of extracting an image including multiple lumbar vertebrae from the spinal image; and A step of extracting an image including one lumbar vertebra from the above spinal image; Integrated analysis method for spinal diseases.

5. In paragraph 2, The step of applying the above spinal segments to the learned disease classification model to determine the type of disease corresponding to the spinal segment is as follows: A step of extracting multiple regions of interest from the above spinal image; A step of determining the type of disease corresponding to the spine in each of the plurality of regions of interest: and A step of determining a type of disease that overlaps with the types of diseases determined in each of the plurality of areas of interest as a type of disease corresponding to the spine; Integrated analysis method for spinal diseases.

6. In paragraph 2, The step of determining the second disease region in the spinal image by applying the above spinal segments to the learned region segmentation model is as follows: A step of extracting multiple regions of interest from the above spinal image; A step of determining a disease area in each of the above multiple areas of interest: and A step of determining an area where the disease areas determined in each of the plurality of areas of interest overlap as a second disease area; Integrated analysis method for spinal diseases.

7. In paragraph 1, The step of applying the above spinal image to the learned spinal segment detection model to detect multiple spinal segments is as follows: A step of detecting each vertebra included in the above spinal image as a bounding box; and A step of detecting the corners of each of the vertebrae in the bounding box; Integrated analysis method for spinal diseases.

8. In paragraph 7, The step of generating measurement indices by measuring the height and width of a plurality of vertebrae and the spacing between the plurality of vertebrae based on the above spinal segments is as follows: A step of measuring the height and width of the plurality of vertebrae and the spacing between the plurality of vertebrae based on the corners of each of the vertebrae; and A step of determining a vertebra or a gap between vertebrae that deviates from the normal standard based on the above measurement indicators; Integrated analysis method for spinal diseases.

9. In paragraph 1, The step of generating integrated disease information of the spine by applying the above measurement indicators and disease information to the learned integrated analysis model is as follows. (i) Based on the above measurement information, the vertebra or the gap between the vertebrae that deviates from the normal standard and (ii) Based on the disease area and disease type determined based on the above imaging disease analysis model, the integrated disease type and integrated disease area are determined. Integrated analysis method for spinal diseases.

10. A computer-readable recording medium having recorded thereon a program for executing a method according to any one of claims 1 to 9 on a computer.

11. In the integrated spinal disease analysis device, A communication unit that receives spinal images and outputs integrated spinal disease information; processor; and memory; including; The above processor, A spine segment detection unit that detects multiple spine segments by applying the above spine image to a learned spine segment detection model; A spine measurement unit that generates measurement indices by measuring the height and width of a plurality of vertebrae and the spacing between the plurality of vertebrae based on the above spinal segments; An imaging disease analysis unit that applies the above spinal segments to one or more learned imaging disease analysis models to generate disease information of the spinal vertebrae; and An integrated analysis unit that applies the above measurement indicators and disease information to a learned integrated analysis model to generate integrated disease information of the spine; Electronic devices containing.

Images