Blood vessel classification method and electronic device

Abstract

To provide a blood vessel classification method and an electronic device achieved by at least one processor.SOLUTION: A blood vessel classification method may include: a step of acquiring a video capturing cardiac blood vessels of a subject and angle information of a video acquisition device with respect to the subject; and a step of identifying the type of at least one cardiac blood vessel contained in the video on the basis of the angle information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present disclosure relates to a blood vessel classification method and an electronic device.

Background Art

[0002] Coronary arteries (or cardiovascular vessels) are arteries surrounding the heart, and can supply oxygen and nutrients to the myocardium, which is the muscle of the heart. That is, the function of the heart can be performed by continuously supplying oxygen and nutrients from the coronary arteries to the myocardium. Accordingly, if an abnormal phenomenon (e.g., a disease) occurs in a coronary artery and oxygen and nutrients are not properly supplied to the myocardium, a cardiovascular disease such as myocardial infarction can develop.

[0003] Such coronary arteries are positioned in a form that surrounds the heart in a coronal shape, and can be divided into detailed types based on their arranged positions. For example, coronary arteries can be divided into the Right Coronary Artery (RCA), which starts from the right side of the entrance of the ascending aorta and mainly runs on the right side of the heart, and the Left Coronary Artery (LCA), which starts from the left side of the entrance of the ascending aorta and mainly runs on the left side of the heart. In addition, the left coronary artery can be more specifically divided into the Left Main Coronary Artery (LMCA, hereinafter referred to as LM) starting from the upper left end of the heart, the Left Anterior Descending coronary artery (LAD) branching from the left main coronary artery, and the Left Circumflex coronary artery (LCX).

[0004] As described above, depending on the positional morphology of the coronary artery, the coronary artery can supply oxygen and nutrients to the adjacent myocardium. Accordingly, when an abnormal phenomenon occurs in a part of the coronary artery, it is possible to determine in which region of the coronary artery the abnormal phenomenon occurs, further identify the myocardial region associated with the region of the coronary artery where the abnormal phenomenon occurs, and diagnose cardiovascular disease more precisely.

[0005] On the other hand, coronary angiography (CAG) may be used to observe whether abnormal phenomena occur in the coronary arteries. CAG refers to an examination in which a contrast agent is injected into the blood vessels of a subject (e.g., a patient), and the coronary arteries are imaged using an X-ray imaging device (e.g., a C-arm X-ray imaging device). The condition of the coronary arteries is then confirmed through the CAG images, and cardiovascular disease is diagnosed.

[0006] Previously, assessing the condition of coronary arteries through CAG imaging relied on the experience of physicians or analysts. However, recently, technologies have been developed that allow electronic devices to receive CAG image input through machine learning and diagnose cardiovascular diseases along with analyzing the CAG images. However, due to the diversity of blood vessel types and morphologies, there is a need to develop technologies to accurately classify blood vessels for more precise analysis. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Korean Registered Patent Publication No. 10-2187842 [Overview of the Initiative] [Problems that the invention aims to solve]

[0008] This disclosure provides a method for classifying blood vessels and an electronic device for solving the aforementioned problems. [Means for solving the problem]

[0009] This disclosure can be embodied in a variety of ways, including methods, apparatus (systems), and / or computer programs.

[0010] According to one embodiment of the present disclosure, a vascular classification method performed by at least one processor may include the steps of acquiring an image including the cardiovascular system of a subject and angle information of an image acquisition device relative to the subject, and identifying at least one type of cardiovascular system included in the image based on the angle information.

[0011] According to one embodiment of the present disclosure, a vascular classification method performed by at least one processor may include the steps of acquiring an image including the cardiovascular system of a subject and identifying at least one type of cardiovascular system included in the image through a machine learning model that takes the image as input.

[0012] According to one embodiment of the present disclosure, the electronic device includes a memory and at least one processor connected to the memory and configured to execute at least one computer-readable program contained in the memory, the at least one program which may include instructions for acquiring images including the cardiovascular system of a subject and angle information of an image acquisition device relative to the subject, and for identifying at least one type of cardiovascular system included in the images based on the angle information.

[0013] According to one embodiment of this disclosure, a computer program can be provided for executing the aforementioned blood vessel classification method on a computer. [Effects of the Invention]

[0014] According to some embodiments of this disclosure, the accuracy of vascular classification can be increased by identifying at least one type of cardiovascular system included in the video footage of the subject's cardiovascular system based on the angle information of the video acquisition device relative to the subject, thereby enabling more precise analysis of the video footage.

[0015] The effects of this disclosure are not limited to those mentioned above, and any other effects not mentioned can be clearly understood by a person with ordinary skill in the art to which this disclosure pertains ("ordinary engineer") from the wording of the claims. [Brief explanation of the drawing]

[0016] Embodiments of the present disclosure are described with reference to the accompanying drawings set forth below, wherein like reference numerals indicate like elements, but the present disclosure is not limited thereto. [Figure 1] It is a drawing exemplarily showing an electronic device for classifying blood vessels according to an embodiment of the present disclosure. [Figure 2] It is a drawing showing an internal structure of an electronic device according to an embodiment of the present disclosure. [Figure 3] It is a perspective view of an electronic device including an image acquisition device according to an embodiment of the present disclosure. [Figure 4] It is a drawing for explaining angle information of an image acquisition device relative to a subject according to an embodiment of the present disclosure. [Figure 5] It is a drawing for explaining a method of classifying blood vessels based on image and angle information according to an embodiment of the present disclosure. [Figure 6] It is a drawing showing information about types of left coronary artery matched with angle information according to an embodiment of the present disclosure. [Figure 7] It is a drawing showing information about types of right coronary artery matched with angle information according to an embodiment of the present disclosure. [Figure 8] It is a drawing for explaining a method of classifying blood vessels according to an embodiment of the present disclosure. [Figure 9] It is a drawing for explaining a method of limiting types of cardiovascular vessels that can be identified in an image based on angle information according to an embodiment of the present disclosure. [Figure 10] It is a drawing for explaining a method of correcting identified types of cardiovascular vessels based on angle information according to an embodiment of the present disclosure. [Figure 11] It is a drawing showing an artificial neural network model according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] <SUMMARY OF THE INVENTION> According to one embodiment, the angle information may include first rotation angle information of an image acquisition apparatus about a first axis in the up-down direction of a subject, and second rotation angle information of the image acquisition apparatus about a second axis in the left-right direction of the subject.

[0018] According to one embodiment, the step of identifying at least one cardiovascular type may include the step of identifying at least one cardiovascular type included in an image based on the angle information and information about a cardiovascular type matched with the angle information.

[0019] According to one embodiment, the machine learning model may be trained to limit the types of cardiovascular vessels that can be identified in an image based on angle information of an image acquisition apparatus with respect to a subject.

[0020] According to one embodiment, the step of identifying at least one cardiovascular type may further include the step of correcting the identified at least one cardiovascular type based on angle information of an image acquisition apparatus with respect to a subject.

[0021] According to one embodiment, the blood vessel classification method further includes the step of training a machine learning model to identify a view corresponding to each of a plurality of images obtained by photographing cardiovascular vessels, based on class information in which view information classified according to angle information of an image acquisition apparatus and a blood vessel type matched with the view information are assigned to one class, with the plurality of images being input, and the step of identifying at least one cardiovascular type may include the step of identifying a view corresponding to an image based on the machine learning model and the step of identifying at least one cardiovascular type included in the image based on the identified view.

[0022] According to one embodiment, the class information may include at least one of a class that enables blood vessel classification for images other than CAG images among the plurality of images, or a dummy class that includes only blood vessel types without view information.

[0023] According to one embodiment, the step of identifying at least one cardiovascular type may further include a step of limiting the cardiovascular types that can be identified in the image based on the identified view.

[0024] According to one embodiment, the step of identifying a view corresponding to a video may include the steps of generating weighted values ​​based on a predetermined number of points in the video, extracting at least one representative frame from the video based on the generated weighted values, inputting at least one representative frame into a machine learning model, and identifying a view corresponding to the video based on the inference results for at least one representative frame through the machine learning model.

[0025] According to one embodiment, the step of identifying a view corresponding to a video based on the inference result for at least one representative frame includes, if at least one representative frame includes multiple representative frames, the step of determining the final result through a voting method for the inference results for multiple representative frames through a machine learning model, the step of determining the final result is, if a valid view is selected as a result of voting for the inference results for multiple representative frames, the step of determining a class including the type of blood vessel and view information corresponding to the selected valid view as the final result, if an invalid view is selected as a result of voting, the step of voting again for the remaining views excluding the invalid view from the inference results for multiple representative frames, and if a valid view is selected as a result of voting again, the step of determining a class including the type of blood vessel and view information corresponding to the selected valid view as the final result. If an invalid view is selected as a result of the vote, the process may include a step in which the remaining frames after removing the invalid view from the inference results for multiple representative frames are voted again, and if all of the results are invalid views, the process is voted again with a second level of confidence, and if a valid view is selected as a result of the vote, the final result is determined to be a class containing the blood vessel type and view information corresponding to the selected valid view. Alternatively, if an invalid view is selected as a result of the vote, the process may include a step in which the remaining frames after removing the invalid view from the inference results for multiple representative frames are voted again, and if all of the results are invalid views, the process is voted again with a second level of confidence, and if the results are either invalid views or images that are not CAG images, the process is voted again with a first level of confidence only for the view information of the class containing the blood vessel type from the initial inference results, and the final result is determined.

[0026] According to one embodiment, the electronic device may include a main body with a built-in lifting drive unit, a lifting unit fixed to the upper end of the lifting drive unit and moving up and down in a first direction, a rotating unit with one end rotatably connected to the lifting unit around a first axis in a second direction perpendicular to the first direction and having a bent surface at the other end, a C-shaped frame unit slidably connected to the bent surface and provided in a "C" shape, an X-ray generator positioned at one end of the C-shaped frame unit, and an image acquisition device positioned at the other end of the C-shaped frame unit.

[0027] According to one embodiment, the angle information may include first rotation angle information of the image acquisition device formed when the rotating part rotates around the first axis, and second rotation angle information of the image acquisition device formed when the C-shaped frame part slides, and centered on a second axis in a third direction perpendicular to the first and second directions.

[0028] <Detailed description of the invention> The specific details for implementing this disclosure will be described below with reference to the attached drawings. However, in the following description, if there is a risk of unnecessarily obscuring the essence of this disclosure, specific descriptions of widely known functions and configurations will be omitted.

[0029] In the attached drawings, identical or corresponding components are given the same reference numerals. Furthermore, in the following description of embodiments, the description of identical or corresponding components may be omitted to avoid duplication. However, the omission of technical details regarding a component does not imply that such a component is not included in any embodiment.

[0030] The advantages and features of the disclosed embodiments, and how they are achieved, will become clear upon reference to the embodiments described below, along with the accompanying drawings. However, this disclosure is not limited to the embodiments disclosed below and may be embodied in a variety of different forms, although these embodiments are provided only to complete the disclosure and to fully inform a person of the ordinary skill of the scope of the invention.

[0031] This specification provides a brief explanation of the terms used and will provide a more detailed explanation in relation to the disclosed embodiments. The terms used herein have been selected as widely used and general terms as possible, taking into account the function of the inventions described herein; however, these may change depending on the intent of the articulates in the relevant field, case law, the emergence of new technologies, etc. In some cases, the applicant has arbitrarily selected terms, in which case their meanings will be described in detail in the description of the relevant invention. Therefore, the terms used in this disclosure should not be simply nouns, but should be defined based on their meaning and the overall content of this disclosure.

[0032] In this specification, singular expressions include plural expressions unless the context clearly identifies them as singular. Similarly, plural expressions include singular expressions unless the context clearly identifies them as plural. When a part of the specification contains any component, this means that it may contain other components, not exclude them, unless otherwise stated.

[0033] Furthermore, the terms “module” or “part” as used in the specification refer to a software or hardware component, and a “module” or “part” performs some role. However, the meaning of “module” or “part” is not limited to software or hardware. A “module” or “part” may be configured to reside on an addressable storage medium, or to regenerate one or more processors. Thus, as an example, a “module” or “part” may include components such as software components, object-oriented software components, class components, and task components, and at least one of the following: processes, functions, attributes, processors, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, or variables. The components and the functions provided within a “module” or “part” may be combined with a smaller number of components and “modules” or further separated into additional components and “modules” or “parts.”

[0034] According to one embodiment of the present disclosure, “module” or “part” may be embodied in a processor and memory. “Processor” should be broadly interpreted to include general-purpose processors, central processing units (CPUs), microprocessors, digital signal processors (DSPs), controllers, microcontrollers, state machines, etc. In some environments, “processor” may refer to application-specific semiconductors (ASICs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), etc. “Processor” may refer to a combination of processing devices such as, for example, a combination of a DSP and a microprocessor, a combination of multiple microprocessors, a combination of one or more microprocessors coupled with a DSP core, or any other combination of such configurations. Also, “memory” should be broadly interpreted to include any electronic component capable of storing electronic information. "Memory" may refer to various types of processor-readable media, such as arbitrary access memory (RAM), read-only memory (ROM), non-volatile arbitrary access memory (NVRAM), programmable read-only memory (PROM), erase-programmable read-only memory (EPROM), electrically erasable PROM (EEPROM), flash memory, magnetic or marking data storage devices, and registers. Memory is said to be in electronic communication with the processor if the processor can read information from and / or record information into it. Memory integrated into a processor is in electronic communication with the processor.

[0035] Furthermore, the terms 1st, 2nd, A, B, (a), (b), etc., used in the following examples are merely used to distinguish one component from another, and do not limit the essence, order, or sequence of the components in question.

[0036] Furthermore, in the following embodiments, when it is stated that a component is “connected,” “joined,” or “connected” to another component, it should be understood that the component may be directly connected to or linked to the other component, but that other components may be further “connected,” “joined,” or “connected” between each component.

[0037] Furthermore, the terms "comprises" and / or "comprising" used in the following embodiments do not preclude the presence or addition of one or more other components, stages, operations, and / or elements mentioned.

[0038] Various embodiments of this disclosure will be described in detail below with reference to the accompanying drawings.

[0039] Figure 1 is an illustrative drawing of an electronic device 100 for classifying blood vessels according to one embodiment of the present disclosure. Referring to Figure 1, the electronic device 100 for classifying blood vessels can provide blood vessel information 120 based on video 112 and angle information 114. For example, the electronic device 100 can acquire video 112 and angle information 114, identify at least one type of blood vessel included in the video 112 based on the angle information 114, and provide blood vessel information 120 for the identified at least one blood vessel. Here, the angle information 114 may represent the angle information of the subject (e.g., patient) of the video acquisition device that captured the video 112.

[0040] In one embodiment, the image 112 may include an image (e.g., a CAG image) of the subject's cardiovascular system (or coronary arteries). For example, the image 112 may include an image of the subject's cardiovascular system taken using an X-ray imaging device (e.g., a C-arm X-ray imaging device) with a contrast agent injected into the subject's blood vessels. In this case, the angle information 114 may indicate the angle information of the image acquisition device included in the X-ray imaging device relative to the subject, and the vascular information 120 may include information for at least one cardiovascular system identified in the image 112. The information for the cardiovascular system may include, for example, the type of cardiovascular system. For the sake of explanation, the following description will describe the case where the image 112 is a CAG image and the angle information 114 is the angle information of the image acquisition device included in the X-ray imaging device relative to the subject, but the type of image 112 and the type of device for acquiring the image 112 are not limited to this.

[0041] In one embodiment, after the subject's cardiovascular system is imaged through the imaging device, the image 112 of the cardiovascular system (e.g., angiography image) can be input to the electronic device 100. For example, the electronic device 100 may be connected to the imaging device by wired or wireless communication, and the image 112 may be provided to the electronic device 100 from the imaging device via a communication module. In one embodiment, the electronic device 100 may be provided integrally with the imaging device. In another example, the electronic device 100 may receive the image 112 from an external electronic device (e.g., an external storage device) connected via a communication module. The method by which the electronic device 100 acquires the image 112 is not limited to the examples described above and may be any method.

[0042] According to one embodiment, angle information 114 of the imaging device (or image acquisition device included in the imaging device) relative to the subject can be input to the electronic device 100 together with the image 112 of the subject's cardiovascular system, or at regular time intervals. For example, if the electronic device 100 is connected to the imaging device by wired or wireless communication, the imaging device can acquire the angle information 114 of the imaging device (or image acquisition device included in the imaging device) relative to the subject at the time the image 112 is captured, and provide the angle information 114 acquired together with the captured image 112 to the electronic device 100 via a communication module. As another example, if the electronic device 100 is provided integrally with the imaging device, the electronic device 100 can acquire the angle information 114 of the imaging device (or image acquisition device included in the imaging device) relative to the subject together with the captured image 112 at the time the imaging device captures the image 112. As yet another example, the electronic device 100 can receive the angle information 114 together with the image 112 from an external electronic device (e.g., an external storage device) connected via a communication module.

[0043] According to one embodiment, the electronic device 100 can identify at least one type of cardiovascular vessel in the video 112 of the subject's cardiovascular system, correct the identified at least one type of cardiovascular vessel based on angle information 114, and then provide vascular information 120 for at least one cardiovascular vessel. For example, after classifying the vessels in the video 112, the electronic device 100 can verify errors in the vascular classification based on angle information 114, and if errors exist in the vascular classification, it can correct the incorrect classification.

[0044] According to one embodiment, the electronic device 100 can limit (or set) the types of cardiovascular vessels that can be identified in the image 112 based on angle information 114, identify at least one type of cardiovascular vessel in the image 112 based on the limited types of cardiovascular vessels, and provide vascular information 120 for the identified at least one cardiovascular vessel. For example, the electronic device 100 can limit (or set) the types of cardiovascular vessels that can be identified in the image 112 based on angle information 114 before classifying the vessels in the image 112. Then, the electronic device 100 can identify at least one type of cardiovascular vessel included in the image 112 so as not to deviate from the limited (or set) types of cardiovascular vessels.

[0045] Figure 2 is a diagram showing the internal configuration of an electronic device 100 according to one embodiment of the present disclosure. Referring to Figure 2, the electronic device 100 may include a memory 210, a processor 220, a communication module 230, and an input / output interface 240. However, the configuration of the electronic device 100 is not limited thereto. According to various embodiments, the electronic device 100 may omit at least one of the aforementioned components and may further include at least one other component. For example, the electronic device 100 may further include a display. In this case, the electronic device 100 can display at least one of the following on the display: a video of a cardiovascular system (e.g., video 112 in Figure 1) or vascular information for a cardiovascular system identified in the video (e.g., vascular information 120 in Figure 1).

[0046] The memory 210 can store a variety of data used by at least one other component of the electronic device 100 (e.g., the processor 220). The data may include, for example, software (or programs) and input or output data for associated instructions.

[0047] Memory 210 can include any non-temporary computer-readable recording medium. In one embodiment, memory 210 can include a permanent mass storage device such as a disk drive, SSD (solid state drive), or flash memory. In another example, a permanent mass storage device such as ROM, SSD, flash memory, or disk drive may be included in the electronic device 100 as a separate permanent storage device distinct from memory 210. Memory 210 can also store an operating system and at least one program code (e.g., instruction words for video analysis and object identification processing installed and driven in the electronic device 100). In Figure 2, memory 210 is shown as a single memory, but this is for illustrative purposes only, and memory 210 can include multiple memories and / or buffer memories.

[0048] Software components may be loaded from a computer-readable recording medium separate from memory 210. Such a separate computer-readable recording medium may include a recording medium that can be directly connected to the electronic device 100, or it may include computer-readable recording media such as a floppy drive, disk, tape, DVD / CD-ROM drive, or memory card. As another example, software components may be loaded into memory 210 via a communication module 230 rather than a computer-readable recording medium. For example, at least one program may be loaded into memory 210 based on a computer program (e.g., a program for data transmission such as contrast-enhanced images of cardiovascular imaging) installed by a file provided through the communication module 230 by a developer or a file distribution system that distributes application installation files.

[0049] The processor 220 can execute software (or programs) to control at least one other component (e.g., hardware or software component) of the electronic device 100 connected to the processor 220, and can perform a variety of data processing or calculations. According to one embodiment, as at least part of the data processing or calculation, the processor 220 can load instructions or data received from other components (e.g., communication module 230) into volatile memory, process the instructions or data stored in volatile memory, and store the resulting data in non-volatile memory.

[0050] The processor 220 may be configured to process instructions for a computer program by performing basic arithmetic, logic, and input / output operations. Instructions may be provided to the electronic device 100 or other external systems via memory 210 or a communication module 230. For example, the processor 220 can identify at least one type of cardiovascular system included in an angiographic image of the cardiovascular system. The processor 220 can then store the vascular information for the identified cardiovascular system in memory 210, output or display it on the display of the electronic device 100, or transmit it to an external electronic device via the communication module 230. Although the processor 220 is shown as a single processor in Figure 2, this is for illustrative purposes only, and the processor 220 may include multiple processors.

[0051] The communication module 230 can assist in establishing a direct (e.g., wired) or wireless communication channel between the electronic device 100 and an external electronic device, and in carrying out communication through the established communication channel. For example, the communication module 230 can provide a configuration or function for the electronic device 100 and an external electronic device (e.g., a user terminal or cloud server) to communicate with each other via a network. As an example, control signals, commands, data, etc., provided by the control of the processor 220 of the electronic device 100 can be transmitted to the external electronic device via the communication module 130 and the network through the communication module of the external electronic device. For example, the electronic device 100 can receive images of the subject's cardiovascular system and angle information of the image acquisition device relative to the subject from the external electronic device via the communication module 230.

[0052] The input / output interface 240 may be a means for interfacing with an input or output device (not shown) that is connected to or may be included in the electronic device 100. For example, the input / output interface 240 may include, but is not limited to, a PCI express interface or an Ethernet interface. In Figure 2, the input / output interface 240 is shown as an element configured separately from the processor 220, but is not limited to this, and the input / output interface 240 may be configured to be included in the processor 220.

[0053] According to one embodiment, the processor 220 can perform functions related to vascular classification. To perform functions related to vascular classification, the processor 220 can execute at least one computer-readable program contained in the memory 210. Here, the at least one program can acquire images of the subject's cardiovascular system (e.g., image 112 in Figure 1) and angle information of the image acquisition device relative to the subject (e.g., angle information 114 in Figure 1), and can include commands for identifying at least one type of cardiovascular system included in the images based on the angle information. Here, the angle information of the image acquisition device relative to the subject can include first rotation angle information of the image acquisition device around a first axis in the vertical direction (or longitudinal direction) of the subject (e.g., the direction connecting the head and feet to each other), and second rotation angle information of the image acquisition device around a second axis in the left-right direction (or width direction) of the subject (e.g., the direction connecting both shoulders or both arms to each other). The angle information will be explained in detail with reference to Figures 3 and 4. Furthermore, for the sake of clarity in the following explanation, the execution of at least one program by the processor 220 to perform functions related to vascular classification can be described as the processor 220 performing functions related to vascular classification. For example, the inclusion of at least one program containing instructions related to functions for vascular classification may be described as the processor 220 performing vascular classification functions.

[0054] According to one embodiment, the processor 220 can identify at least one type of cardiovascular vessel included in the video based on angle information and information about the type of cardiovascular vessel matched with the angle information. The information about the type of cardiovascular vessel matched with the angle information may be, for example, information that defines (or sets) the types of cardiovascular vessels that can be identified in the captured video based on the shooting angle of the video acquisition device. According to one embodiment, the angle information and the information about the type of cardiovascular vessel matched with the angle information may be pre-stored in the memory 210 in a table data structure. The information about the angle information and the information about the type of cardiovascular vessel matched with the angle information will be described in detail with reference to Figures 6 and 7.

[0055] According to one embodiment, the processor 220 can identify at least one cardiovascular type contained in a video through a machine learning model that takes the video as input. Here, the machine learning model may be contained in any memory accessible to the processor 220 (e.g., memory 210). The machine learning model may also include any model used to infer a solution for a given input. According to one embodiment, the machine learning model may include an artificial neural network model that includes an input layer, a plurality of hidden layers, and an output layer. Here, each layer may contain one or more nodes. The machine learning model may also include weights associated with the plurality of nodes contained in the machine learning model. Here, the weights may include any parameters associated with the machine learning model. The machine learning model in this disclosure may be a model that has been trained using a variety of learning methods. For example, a variety of learning methods such as supervised learning, semi-supervised learning, unsupervised learning (or autonomous learning), and reinforcement learning may be used in this disclosure. In this disclosure, the machine learning model may refer to an artificial neural network model, and the artificial neural network model may refer to a machine learning model. Artificial neural network models will be explained in detail with reference to Figure 11.

[0056] Figure 3 is a perspective view of an electronic device 300 including an image acquisition device according to one embodiment of the present disclosure. The electronic device 300 shown in Figure 3 is a device including an image acquisition device 352 for capturing images (e.g., image 112 in Figure 1) of the cardiovascular system of a subject, and may be connected to the electronic device 100 in Figure 1 via a communication module or provided integrally with the electronic device 100 in Figure 1. The electronic device 300 may be, for example, an X-ray imaging device.

[0057] Referring to Figure 3, the electronic device 300 may include a main body 310, a lifting unit 320, a rotating unit 330, a C-shaped frame unit 340, an X-ray generator 354, and an image acquisition device 352. However, the configuration of the electronic device 300 is not limited thereto. The electronic device 300 may be provided in any configuration and form as long as it includes the image acquisition device 352.

[0058] The main body 310 may have a built-in lifting drive unit. In one embodiment, the main body 310 may have a built-in control unit (e.g., a processor) for controlling the configuration of the electronic device 300. Also, if the electronic device 300 is connected to the electronic device 100 in Figure 1 via a communication module, the main body 310 may have a built-in communication module.

[0059] The lifting unit 320 is fixed to the upper end of a lifting drive unit built into the main body 310 and can be raised and lowered in a first direction 392 (e.g., vertical direction). For example, the lifting unit 320 can be adjusted to match the height of the object being examined (or photographed) (e.g., the heart) depending on the posture of the subject.

[0060] The rotating part 330 is connected to the lifting part 320 at one end so as to be rotatable 394 about a first axis in a second direction perpendicular to the first direction 392, and a bent surface may be formed at the other end. When the rotating part 330 rotates 394 about the first axis, the rotation angle (e.g., first rotation angle) of the image acquisition device 352 may be changed. Here, the first axis may be an axis in the vertical direction (or longitudinal direction) (e.g., the direction in which the head and feet are connected to each other) or a parallel axis of the subject. For example, the rotation 394 of the rotating part 330 about the first axis indicates that the image acquisition device 352 rotates about the first axis, and the rotation of the image acquisition device 352 about the first axis indicates that the image acquisition device 352 rotates left and right around the torso (or heart) of the subject. In the following description, the rotation angle information of the image acquisition device 352 formed when the rotating part 330 rotates 394 about the first axis may be referred to as the first rotation angle information.

[0061] The C-shaped frame section 340 is slidably connected to a bent surface formed on the rotating section 330 and may be provided in a "C" shape (or a ring shape with a portion cut off). Depending on the shape of such a C-shaped frame section 340, the electronic device 300 may be referred to as a C-arm or C-arm imaging device. When the C-shaped frame section 340 slides 396 on the bent surface formed on the rotating section 330, the rotation angle (e.g., second rotation angle) of the image acquisition device 352 may be changed. When the C-shaped frame section 340 slides 396 on the bent surface formed on the rotating section 330, the image acquisition device 352 can rotate about a second axis in a third direction perpendicular to the first direction 392 and the second direction. Here, the second axis may be an axis in the left-right direction (or width direction) of the subject (e.g., the direction in which both shoulders or both arms are connected to each other) or a parallel axis. For example, the sliding 396 of the C-shaped frame portion 340 on the bent surface formed on the rotating portion 330 indicates that the image acquisition device 352 rotates around the second axis, and the rotation of the image acquisition device 352 around the second axis indicates that the image acquisition device 352 rotates vertically around the torso (or heart) of the subject. In the following description, the rotation angle information of the image acquisition device 352 formed when the C-shaped frame portion 340 slides 396 on the bent surface formed on the rotating portion 330 may be referred to as the second rotation angle information.

[0062] The X-ray generator 354 is positioned at one end of the C-shaped frame section 340, and the image acquisition device 352 may be positioned at the other end of the C-shaped frame section 340. The X-ray generator 354 can generate X-rays and transmit them through the object to be inspected (or the object to be photographed), and the amount of transmitted X-rays can be detected by the image acquisition device 352, which can then process this signal to acquire an image (e.g., image 112 in Figure 1). At this time, the electronic device 300 can acquire the image acquired through the image acquisition device 352, along with the angle information of the image acquisition device 352 at the time the image was taken (e.g., first rotation angle information and second rotation angle information) (e.g., angle information 114 in Figure 1).

[0063] Figure 4 is a diagram illustrating the angular information of the image acquisition device 352 for a subject according to one embodiment of the present disclosure. Referring to Figure 4, the X-ray imaging apparatus (or C-arm or C-arm imaging apparatus) (e.g., the electronic device 300 in Figure 3) can change (or set) the angles 422 and 424 of the image acquisition device 352 to correspond to the cardiovascular region to be imaged in order to confirm which region of the subject's cardiovascular system has an abnormal phenomenon. As described with reference to Figure 3, the X-ray imaging apparatus has an X-ray generator 354 and an image acquisition device 352 positioned at both ends of a C-shaped frame section 340. The angles 422 and 424 of the image acquisition device 352 can be changed by the C-shaped frame section 340 sliding on the curved surface of the rotating section (e.g., the rotating section 330 in Figure 3) (e.g., sliding 396 in Figure 3) or by the rotating section rotating (e.g., rotation 394 in Figure 3).

[0064] The angles 422 and 424 of the image acquisition device 352 can be set relative to the subject of examination (or the subject being photographed). For example, when the subject lies on the examination table 410, the vertical direction (or longitudinal direction) of the subject (e.g., the direction connecting the head and feet) may be the X-axis direction, and the left-right direction (or width direction) of the subject (e.g., the direction connecting both shoulders or both arms) may be the Y-axis direction. More specifically, the direction from the head to the feet may be the (+)X-axis direction, the direction from the feet to the head may be the (-)X-axis direction, the direction from the right shoulder (or right arm) to the left shoulder (or left arm) may be the (+)Y-axis direction, and the direction from the left shoulder (or left arm) to the right shoulder (or right arm) may be the (-)Y-axis direction. At this time, the first rotation angle information of the image acquisition device 352, which is formed when the rotating part rotates and the image acquisition device 352 rotates around the X axis, can be set as the first rotation angle (α) 422, and the second rotation angle information of the image acquisition device 352, which is formed when the C-shaped frame part 340 slides on the bent surface of the rotating part and the image acquisition device 352 rotates around the Y axis, can be set as the second rotation angle (β) 424.

[0065] The first rotation angle 422 can be referred to as the primary angle. When the first rotation angle 422 has a rotation angle in the (-)Y-axis direction, the image can be referred to as having an RAO view, and the first rotation angle 422 can be expressed as the RAO angle. Also, when the first rotation angle 422 has a rotation angle in the (+)Y-axis direction, the image can be referred to as having an LAO view, and the first rotation angle 422 can be expressed as the LAO angle. Furthermore, when the first rotation angle 422 has a rotation angle centered on the Y-axis (i.e., 0 degrees), the image can be referred to as having an AP view.

[0066] The second rotation angle 424 may be referred to as the secondary angle. When the second rotation angle 424 has a rotation angle in the (-)X-axis direction, the image may be referred to as having a CRA (or CRANIAL) view, and the second rotation angle 424 may be expressed as the CRA angle. Also, when the second rotation angle 424 has a rotation angle in the (+)X-axis direction, the image may be referred to as having a CAU (or CAUDAL) view, and the second rotation angle 424 may be expressed as the CAU angle.

[0067] According to one embodiment, the type of cardiovascular system to be observed or included in the video can be set through a combination of a first rotation angle 422 and a second rotation angle 424. The types of cardiovascular systems that can be set through a combination of a first rotation angle 422 and a second rotation angle 424 will be explained in detail with reference to Figures 6 and 7.

[0068] Figure 5 is a diagram illustrating a method for classifying blood vessels based on image and angle information according to one embodiment of the present disclosure. An X-ray imaging device (or C-arm or C-arm imaging device) (e.g., electronic device 300 in Figure 3) can change (or set) the angle of the image acquisition device (e.g., image acquisition device 352 in Figures 3 and 4) to correspond to the cardiovascular region to be imaged in order to determine which region of the subject's cardiovascular system has experienced an abnormal phenomenon. The X-ray imaging device can also transmit the captured image (e.g., CAG image) and the angle information of the image acquisition device at the time of imaging to an electronic device (e.g., electronic device 100 in Figures 1 and 2) once the image acquisition device has imaged the subject's cardiovascular system. Here, the X-ray imaging device (e.g., electronic device 300 in Figure 3) and the electronic device that provides vascular information via images (e.g., electronic device 100 in Figures 1 and 2) are provided as a single unit or can operate in conjunction with each other. Therefore, for the sake of explanation, in the following description, they will be described as an integrated electronic device.

[0069] Referring to Figure 5, the electronic device can identify at least one cardiovascular type contained in the images 512 and 532 based on the acquired images 512 and 532 and the angle information 514 and 534 of the image acquisition device relative to the subject at the time the images 512 and 532 were captured, and provide vascular information 520 and 540 for the identified at least one cardiovascular. For example, the electronic device can identify at least one cardiovascular type contained in the first image 512 based on the first image 512 and the first angle information 514 of the image acquisition device relative to the subject at the time the first image 512 was captured, and provide first vascular information 520 for the identified at least one cardiovascular. Furthermore, the electronic device can identify at least one cardiovascular type contained in the second image 532 based on the second angle information 534 of the image acquisition device relative to the subject at the time the second image 532 was captured, and provide second vascular information 540 for the identified at least one cardiovascular. However, the number of images that can be used for vascular classification and the angle information of the image acquisition device relative to the subject at the time the images were taken are not limited to these. For example, although not shown in the figures, the electronic device can identify at least one cardiovascular type contained in at least one image based on at least one image different from the first image 512 and the second image 532 and the angle information of the image acquisition device relative to the subject at the time the images were taken, and provide vascular information for the identified at least one cardiovascular.

[0070] According to one embodiment, the electronic device can identify at least one cardiovascular type in the images 512 and 532, and then correct the identified at least one cardiovascular type based on the angle information 514 and 534. Subsequently, the electronic device can provide vascular information 520 and 540 for at least one cardiovascular. For example, after classifying the blood vessels in the images 512 and 532, the electronic device can verify errors in the vascular classification based on the angle information 514 and 534, and if errors exist in the vascular classification, it can correct the errors.

[0071] According to one embodiment, the electronic device can, after limiting (or setting) the types of cardiovascular vessels that can be identified in the images 512, 532 based on angular information 514, 534, identify at least one type of cardiovascular vessel in the images 512, 532 based on the limited types of cardiovascular vessels. Subsequently, the electronic device can provide vascular information 520, 540 for the identified at least one cardiovascular vessel. For example, before classifying the vessels in the images 512, 532, the electronic device can limit (or set) the types of cardiovascular vessels that can be identified in the images 512, 532 based on angular information 514, 534, and identify at least one type of cardiovascular vessel included in the images 512, 532 so as not to deviate from the limited (or set) types of cardiovascular vessels.

[0072] Figure 6 is a diagram showing information regarding the type of left coronary artery matched with angle information according to one embodiment of the present disclosure. An X-ray imaging device (or C-arm or C-arm imaging device) (e.g., electronic device 300 in Figure 3) can set different cardiovascular types included in the image (e.g., CAG image) depending on the angle (e.g., angles 422 and 424 in Figure 4) of the image acquisition device (e.g., image acquisition device 352 in Figures 3 and 4). Figure 6 will illustrate the types of left coronary arteries that can be set through a combination of the first rotation angle (α) and the second rotation angle (β) of the image acquisition device. In Figure 6, the angle information of the image acquisition device and the information regarding the type of left coronary artery matched with the angle information are shown as having a table data structure, but the data structure is not limited to this.

[0073] Referring to Figure 6, when the first rotation angle (α) has an RAO angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (-)Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (-)X axis direction, the image can have an RAO CRANIAL view. When the image has an RAO CRANIAL view, the left anterior descending coronary artery (LAD) can be identified in the image. In the table in Figure 6, the "+" symbol indicates that identification is possible, the "++" symbol indicates that identification is good, and the "-" symbol indicates that identification is difficult. For example, when the identification degrees indicated by the "+", "++", and "-" symbols are set as the first identification degree, second identification degree, and third identification degree, respectively, the second identification degree may have the largest value, the third identification degree may have the smallest value, and the first identification degree may have a value between the second and third identification degrees. Furthermore, in the table in Figure 6, "LAD-p" can represent the proximal portion of the left anterior descending coronary artery, "LAD-m" can represent the middle portion of the left anterior descending coronary artery, and "LAD-d" can represent the distal portion of the left anterior descending coronary artery. Also, in the table in Figure 6, "LCX-p" can represent the proximal portion of the left circumflex coronary artery, and "LCX-d" can represent the distal portion of the left circumflex coronary artery. For example, when the image has a RAO CRANIAL view, the middle and distal portions of the left anterior descending coronary artery are best identified in the image, the proximal portion of the left anterior descending coronary artery is identifiable, and the proximal and distal portions of the left circumflex coronary artery may be difficult to identify.

[0074] When the first rotation angle (α) has an AP angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device is 0 degrees and the second rotation angle of the image acquisition device has a rotation angle in the (-)X axis direction, the image can have an AP CRANIAL view. When the image has an AP CRANIAL view, the proximal, intermediate, and distal portions of the left anterior descending coronary artery can be identified in the image, while the proximal and distal portions of the left circumflex coronary artery may be difficult to identify.

[0075] When the first rotation angle (α) has an LAO angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (+) Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (-) X axis direction, the image can have an LAO CRANIAL view. When the image has an LAO CRANIAL view, the proximal, intermediate, and distal portions of the left main coronary artery (LM) and the left anterior descending coronary artery can be identified in the image, while the proximal and distal portions of the left circumflex coronary artery may be difficult to identify.

[0076] When the first rotation angle (α) has an AP angle and the second rotation angle (β) is 0 degrees, that is, when the first rotation angle of the image acquisition device is 0 degrees and the second rotation angle of the image acquisition device is also 0 degrees, the image can have an AP view. When the image has an AP view, the left anterior descending coronary artery may be identifiable in the image.

[0077] When the first rotation angle (α) has an RAO angle and the second rotation angle (β) has a CAU angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (-)Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (+)X axis direction, the image can have an RAO CAUDAL view. When the image has an RAO CAUDAL view, the left main coronary artery, the proximal part of the left anterior descending coronary artery, and the proximal and distal parts of the left circumflex coronary artery can be identified in the image, while the intermediate and distal parts of the left anterior descending coronary artery may be difficult to identify.

[0078] When the first rotation angle (α) has an AP angle and the second rotation angle (β) has a CAU angle, that is, when the first rotation angle of the image acquisition device is 0 degrees and the second rotation angle of the image acquisition device has a rotation angle in the (+)X axis direction, the image can have an AP CAUDAL view. When the image has an AP CAUDAL view, the left main coronary artery is best identified in the image, the proximal part of the left anterior descending coronary artery, the proximal and distal parts of the left circumflex coronary artery are identifiable, and the middle and distal parts of the left anterior descending coronary artery may be difficult to identify.

[0079] When the first rotation angle (α) has an LAO angle and the second rotation angle (β) has a CAU angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (+)Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (+)X axis direction, the image can have an LAO CAUDAL view. The LAO CAUDAL view may be referred to as a SPIDER view. When the image has an LAO CAUDAL view, the left main coronary artery is best identified in the image, the proximal part of the left anterior descending coronary artery and the proximal part of the left circumflex coronary artery are identifiable, and the middle and distal parts of the left anterior descending coronary artery and the distal part of the left circumflex coronary artery may be difficult to identify.

[0080] According to one embodiment, the electronic device (e.g., the electronic device 100 in Figures 1 and 2) can process the video so that cardiovascular types that are difficult to identify (e.g., indicated by a - symbol) according to the identification level set in the aforementioned table are not identified. For example, the electronic device can process (or restrict) the left circumflex coronary artery so that it is not identified in a video with a RAO CRANIAL view.

[0081] Figure 7 is a diagram showing information regarding the type of right coronary artery matched with angle information according to one embodiment of the present disclosure. Referring to Figure 7, an X-ray imaging device (or C-arm or C-arm imaging device) (e.g., electronic device 300 in Figure 3) can set different cardiovascular types included in the image (e.g., CAG image) depending on the angle (e.g., angles 422, 424 in Figure 4) of the image acquisition device (e.g., image acquisition device 352 in Figures 3 and 4). In Figure 7, we will describe the types of right coronary arteries that can be set through a combination of the first rotation angle (α) and the second rotation angle (β) of the image acquisition device. In Figure 7, the angle information of the image acquisition device and the information regarding the type of right coronary artery matched with the angle information are shown as having a table data structure, but the data structure is not limited to this.

[0082] Referring to Figure 7, when the first rotation angle (α) has an RAO angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (-)Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (-)X axis direction, the image can have an RAO CRANIAL view. In the table in Figure 7, the "+" symbol indicates identifiable, the "++" symbol indicates good identifiable, and the "-" symbol indicates difficult identifiable. Also, in the table in Figure 7, "prox." indicates the proximal part, "mid." indicates the middle part, and "dis." indicates the distal part. For example, when the image has an RAO CRANIAL view, the proximal part of the right coronary artery is best identified in the image, and the middle part of the right coronary artery may be identifiable.

[0083] When the first rotation angle (α) has an AP angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device is 0 degrees and the second rotation angle of the image acquisition device has a rotation angle in the (-)X axis direction, the image can have an AP CRANIAL view. When the image has an AP CRANIAL view, the distal portion of the right coronary artery is best identified in the image, and the middle portion of the right coronary artery may be identifiable.

[0084] When the first rotation angle (α) has an LAO angle and the second rotation angle (β) has a CRA angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (+) Y-axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (-) X-axis direction, the image can have an LAO CRANIAL view. When the image has an LAO CRANIAL view, the proximal portion of the right coronary artery is best identified in the image, and the middle and distal portions of the right coronary artery may be identifiable.

[0085] When the first rotation angle (α) has an RAO angle and the second rotation angle (β) is 0 degrees, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (-)Y axis direction and the second rotation angle of the image acquisition device is 0 degrees, the image can have an RAO view. When the image has an RAO view, the middle portion of the right coronary artery is best identified in the image, while the proximal and distal portions of the right coronary artery may be difficult to identify.

[0086] When the first rotation angle (α) has an LAO angle and the second rotation angle (β) is 0 degrees, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (+)Y axis direction and the second rotation angle of the image acquisition device is 0 degrees, the image can have an LAO view. When the image has an LAO view, the proximal, intermediate, and distal portions of the right coronary artery may be distinguishable in the image.

[0087] When the first rotation angle (α) has an LAO angle and the second rotation angle (β) has a CAU angle, that is, when the first rotation angle of the image acquisition device has a rotation angle in the (+)Y axis direction and the second rotation angle of the image acquisition device has a rotation angle in the (+)X axis direction, the image can have an LAO CAUDAL view. When the image has an LAO CAUDAL view, the proximal and intermediate portions of the right coronary artery may be distinguishable in the image.

[0088] According to one embodiment, the electronic device (e.g., the electronic device 100 in Figures 1 and 2) can process the video so that cardiovascular types that are difficult to identify (e.g., indicated by a - symbol) according to the identification level set in the aforementioned table are not identified. For example, the electronic device can process (or restrict) the identification of the proximal and distal portions of the right coronary artery in a video with a RAO view.

[0089] Figure 8 is a diagram illustrating a method for classifying blood vessels according to one embodiment of the present disclosure. Referring to Figure 8, the processor (e.g., processor 220 in Figure 2) of an electronic device for classifying blood vessels (e.g., electronic device 100 in Figures 1 and 2) can acquire, in 810 steps (S810), video footage of the subject's cardiovascular system (e.g., video 112 in Figure 1) and angle information (e.g., angle information 114 in Figure 1 or angle information 422, 424 in Figure 4) from an image acquisition device (e.g., image acquisition device 352 in Figures 3 and 4) on the subject. Here, the angular information of the image acquisition device relative to the subject may include first rotational angle information of the image acquisition device around the first axis (e.g., the X-axis in Figure 4) in the vertical direction (or longitudinal direction) of the subject (e.g., the direction connecting the head and feet to each other), and second rotational angle information of the image acquisition device around the second axis (e.g., the Y-axis in Figure 4) in the left-right direction (or width direction) of the subject (e.g., the direction connecting both shoulders or both arms to each other). According to one embodiment, the processor can acquire image and angular information simultaneously or at regular time intervals.

[0090] In step 820 (S820), the processor can identify at least one cardiovascular type included in the image based on the angle information. According to one embodiment, the processor can identify at least one cardiovascular type included in the image based on the angle information and information about the cardiovascular type matched with the angle information. Here, the information about the cardiovascular type matched with the angle information may be information that defines (or sets) the cardiovascular types that can be identified in the captured image depending on the shooting angle of the image acquisition device. According to one embodiment, the information about the angle information and information about the cardiovascular type matched with the angle information is a table data structure that can be pre-stored in the memory of the electronic device (e.g., memory 210 in Figure 2).

[0091] According to one embodiment, after the processor identifies at least one cardiovascular type in the video, it can correct the identified at least one cardiovascular type based on the angle information. For example, after classifying the blood vessels in the video, the processor can verify errors in the blood vessel classification based on the angle information, and if errors exist in the blood vessel classification, it can correct the incorrect classification.

[0092] According to one embodiment, the processor can limit (or set) the types of cardiovascular vessels that can be identified in the image based on angular information, and can identify at least one type of cardiovascular vessel in the image based on the limited types of cardiovascular vessels. For example, before classifying the vessels in the image, the processor can limit (or set) the types of cardiovascular vessels that can be identified in the image based on angular information, and can identify at least one type of cardiovascular vessel included in the image so as not to deviate from the limited (or set) types of cardiovascular vessels.

[0093] Figure 9 is a diagram illustrating a method for limiting the types of cardiovascular vessels that can be identified in a video based on angle information according to one embodiment of the present disclosure. A processor (e.g., processor 220 in Figure 2) of an electronic device for classifying blood vessels (e.g., electronic device 100 in Figures 1 and 2) can identify at least one type of cardiovascular vessel included in a video based on a video of the subject's cardiovascular system (e.g., video 112 in Figure 1) and angle information (e.g., angle information 114 in Figure 1 or angle information 422, 424 in Figure 4) from a video acquisition device (e.g., video acquisition device 352 in Figures 3 and 4) on the subject. In this process, the processor can identify at least one type of cardiovascular vessel included in the video through a machine learning model (e.g., a vascular classification model) that takes the video as input.

[0094] Referring to Figure 9, at step 910 (S910), the processor can input angular information into the machine learning model to limit the types of cardiovascular tissue that can be identified in the image. For example, the machine learning model can be trained to limit the types of cardiovascular tissue that can be identified in the image based on the angular information when additional angular information is input into the image.

[0095] At 920 steps (S920), the processor can identify at least one cardiovascular type in the video through a machine learning model. For example, the machine learning model can identify at least one cardiovascular type in the video based on a limited set of cardiovascular types. That is, the machine learning model can identify at least one cardiovascular type in the video without deviating from a limited (or set) set of cardiovascular types.

[0096] According to one embodiment, the machine learning model used in steps 910 and 920 may be the same machine learning model. For example, the machine learning model may include one that can receive video and angle information inputs simultaneously to narrow down the types of blood vessels and classify blood vessels in video.

[0097] According to one embodiment, the machine learning models used in the 910-step and 920-step processes may be different from each other. For example, the machine learning model used in the 910-step process may include a machine learning model that receives angle information as input and outputs information about the types of blood vessels that are limited in identification in the video, while the machine learning model used in the 920-step process may include a machine learning model that receives video and information about the types of blood vessels that are limited in identification in the video as input and can classify blood vessels in the video.

[0098] Figure 10 is a diagram illustrating a method for correcting cardiovascular types identified based on angle information according to one embodiment of the present disclosure. A processor (e.g., processor 220 in Figure 2) of an electronic device for classifying blood vessels (e.g., electronic device 100 in Figures 1 and 2) can identify at least one cardiovascular type included in the video based on the video footage of the subject's cardiovascular system (e.g., video 112 in Figure 1) and angle information (e.g., angle information 114 in Figure 1 or angle information 422, 424 in Figure 4) from an image acquisition device (e.g., image acquisition device 352 in Figures 3 and 4) on the subject. In this process, the processor can identify at least one cardiovascular type included in the video through a machine learning model (e.g., a vascular classification model) that takes the video as input.

[0099] Referring to Figure 10, in step 1010 (S1010), the processor can identify at least one cardiovascular type contained in the video through a machine learning model. For example, the machine learning model can identify at least one cardiovascular type in the input video.

[0100] In step 1020 (S1020), the processor can correct at least one identified cardiovascular type based on angular information. For example, the processor can correct at least one identified cardiovascular type based on angular information. According to one embodiment, the processor can correct the identified cardiovascular type based on angular information through a machine learning model. In this case, the machine learning model that can be used in step 1020 may be the same machine learning model used in step 1010. For example, the machine learning model may include a model that simultaneously receives video and angular information input, classifies blood vessels in the video, verifies errors in the blood vessel classification, and corrects errors if errors exist in the blood vessel classification. Alternatively, the machine learning model that can be used in step 1020 may be a different machine learning model from the one used in step 1010. For example, the machine learning model used in step 1010 includes a machine learning model that receives video input and classifies blood vessels in the video, while the machine learning model that can be used in step 1020 may include a model that receives blood vessel classification information (e.g., video with classified blood vessels) and angular information input, verifies errors in the blood vessel classification, and corrects errors if errors exist in the blood vessel classification.

[0101] According to one embodiment, the processor can classify the views that an image can have based on angular information. Here, the angular information may include a first rotation angle and a second rotation angle of the image acquisition device, where the first rotation angle indicates the rotation angle of the subject in the left-right direction (or width direction) (e.g., the direction connecting both shoulders or both arms) (hereinafter referred to as the Y-axis direction), and the second rotation angle indicates the rotation angle of the subject in the up-down direction (or longitudinal direction) (e.g., the direction connecting the head and feet) (hereinafter referred to as the X-axis direction). For example, in the case of the right coronary artery, the processor can classify it as an AP CRANIAL view when the first rotation angle included in the angular information is 0 degrees and the second rotation angle is a rotation angle in the (-)X-axis direction. Also, in the case of the right coronary artery, the processor can classify it as an LAO CRANIAL view when the first rotation angle included in the angular information is a rotation angle in the (+)Y-axis direction and the second rotation angle is a rotation angle in the (-X) axis direction. Furthermore, in the case of the right coronary artery, the processor can classify it as an LAO view when the first rotation angle included in the angle information is a rotation angle in the (+) Y-axis direction and the second rotation angle is 0 degrees. Furthermore, in the case of the right coronary artery, the processor can classify it as an RAO view when the first rotation angle included in the angle information is a rotation angle in the (-) Y-axis direction and the second rotation angle is 0 degrees. Furthermore, in the case of the left anterior descending coronary artery, the processor can classify it as an AP view when the first rotation angle included in the angle information is 0 degrees and the second rotation angle is also 0 degrees. Furthermore, in the case of the left anterior descending coronary artery, the processor can classify it as an AP CRANIAL view when the first rotation angle included in the angle information is 0 degrees and the second rotation angle is a rotation angle in the (-) X-axis direction. Furthermore, in the case of the left anterior descending coronary artery, the processor can classify it as an LAO CRANIAL view when the first rotation angle included in the angle information is a rotation angle in the (+) Y-axis direction and the second rotation angle is a rotation angle in the (-) X-axis direction. Furthermore, in the case of the left anterior descending coronary artery, the processor can classify it as an RAO CRANIAL view when the first rotation angle included in the angle information is a rotation angle in the (-)Y axis direction and the second rotation angle is a rotation angle in the (-)X axis direction.Furthermore, in the case of a left circumflex coronary artery, the processor can classify it as an AP view when the first rotation angle included in the angle information is 0 degrees and the second rotation angle is also 0 degrees. Also, in the case of a left circumflex coronary artery, the processor can classify it as an AP CAUDAL view when the first rotation angle included in the angle information is 0 degrees and the second rotation angle is a rotation angle in the (+)X axis direction. Also, in the case of a left circumflex coronary artery, the processor can classify it as an RAO CAUDAL view when the first rotation angle included in the angle information is a rotation angle in the (-)Y axis direction and the second rotation angle is a rotation angle in the (+)X axis direction. Also, in the case of a left main coronary artery, the processor can classify it as an LAO CAUDAL view when the first rotation angle included in the angle information is a rotation angle in the (+)Y axis direction and the second rotation angle is a rotation angle in the (+)X axis direction.

[0102] Subsequently, the processor can specify (or assign) the classified view information to a class along with the vessel type, as shown in Table 1. At this time, the processor can add classes to the class list that enable vascular classification even for images other than CAG images among the cardiovascular images (e.g., the ("Non-CAG", "")) class in Table 1). In addition, the processor can add dummy classes (e.g., the ("RCA", "") class, the ("LAD", "") class, the ("LCX", "") class and the ("LM", "") class in Table 1) that contain only the vessel type without view information, so that it can handle cases where the acquisition angle is located among many views and it is difficult to classify it as any one of them.

[0103] [Table 1] In the following explanation, as shown in Table 1, view information specified in a class along with the blood vessel type may be referred to as a valid view, while combinations of blood vessel types and view information specified in a dummy class or not specified in a class (e.g., a combination of RCA and AP views) may be referred to as invalid views.

[0104] Subsequently, the processor can train a machine learning model to classify views using video as input based on class information (e.g., a class list). At this time, the processor can limit the types of major blood vessels that can be identified for each view and train the machine learning model to classify the major blood vessels in the video as well. In one embodiment, the processor can process the identification numbers of classes associated with AP views (e.g., the "LAD", "AP" classes and the "LCX", "AP" classes in Table 1) as identical, as shown in the class information in Table 1. In this case, the processor can train the machine learning model so that the types of blood vessels identifiable in an AP view can be at least one of the left anterior descending coronary artery or the left circumflex coronary artery. In another embodiment, the processor can set the identification numbers of classes associated with AP views (e.g., the "LAD", "AP" classes and the "LCX", "AP" classes in Table 1) to be different from each other. In this case, the processor can train the machine learning model so that the types of blood vessels identifiable in an AP view can be the left anterior descending coronary artery or the left circumflex coronary artery.

[0105] Subsequently, the processor can input the video into the machine learning model in connection with the inference process through the trained machine learning model. According to one embodiment, the processor can generate weights (e.g., Gaussian weights) based on a predetermined number of points in the video based on the intensity of the contrast agent, and generate (or extract) at least one representative frame based on the generated weights. Alternatively, the processor can generate weights based on a fixed predetermined number of points, and generate (or extract) at least one representative frame based on the generated weights. Subsequently, the processor can input at least one representative frame into the machine learning model.

[0106] Subsequently, the processor can classify the views based on the inference results for at least one representative frame through a machine learning model. At this point, the processor can classify the views based on class information. The processor can also limit the types of blood vessels that can be identified in the video based on the classified views. Subsequently, the processor can classify the blood vessels contained in the video.

[0107] According to one embodiment, when a processor classifies views based on inference results for multiple representative frames through a machine learning model, it can determine the final result based on class information. For example, the processor can determine the final result for inference results for multiple representative frames through a voting mechanism. More specifically, if a valid view is selected as a result of voting on the inference results for multiple representative frames based on class information, the processor can determine the combination of blood vessel type and view information corresponding to the selected valid view as the final result. Alternatively, if an invalid view is selected as a result of voting on the inference results for multiple representative frames based on class information, the processor can vote again on the remaining views excluding the invalid view. In this case, if a valid view is selected as a result of voting again, the processor can determine the combination of blood vessel type and view information corresponding to the selected valid view as the final result. Alternatively, if all views are invalid as a result of voting again, the processor can vote again with a second confidence level. Subsequently, if a valid view is selected as a result of voting again, the processor can determine the combination of blood vessel type and view information corresponding to the selected valid view as the final result. Alternatively, if the processor votes again and the result is an invalid view or a non-CAG image (e.g., Non-CAG), it can vote again with first confidence only for the view information of the class that contains the blood vessel type from the initial inference result, and then determine the final result.

[0108] Figure 11 is a diagram showing an artificial neural network model 1100 according to one embodiment of the present disclosure. Referring to Figure 11, the artificial neural network model 1100 is an example of a machine learning model and can represent a statistical learning algorithm or a structure that executes such an algorithm, which is embodied in machine learning technology and cognitive science based on the structure of a biological neural network.

[0109] According to one embodiment, the artificial neural network model 1100 can demonstrate a machine learning model with problem-solving capabilities by having nodes, which are artificial neurons that form a network through synaptic connections, learn to repeatedly adjust the synaptic weights so that the error between the correct output corresponding to a specific input and the inferred output decreases. For example, the artificial neural network model 1100 can include any probabilistic model or neural network model used in artificial intelligence learning methods such as machine learning and deep learning.

[0110] According to one embodiment, the aforementioned vascular classification model can be generated in the form of an artificial neural network model 1100. For example, the artificial neural network model 1100 can receive video footage of the subject's cardiovascular system and angle information of the video acquisition device relative to the subject, and based on this, can estimate the type of at least one cardiovascular system included in the video.

[0111] The artificial neural network model 1100 can be implemented as a multi-layer perceptron (MLP) composed of multiple layers of nodes and connections between them. The artificial neural network model 1100 according to this embodiment can be implemented using one of the artificial neural network model structures, including a multi-layer perceptron. The artificial neural network model 1100 can consist of an input layer 1120 that receives input data 1110 (or input signal) from the outside, an output layer 1140 that outputs output data 1150 (or output signal) corresponding to the input data 1110, and n hidden layers 1130_1 to 1130_n (where n is an integer of quantity) located between the input layer 1120 and the output layer 1140, which receive signals from the input layer 1120, extract characteristics, and transmit them to the output layer 1140. Here, the output layer 1140 can receive signals from the hidden layers 1130_1 to 1130_n and output them to the outside.

[0112] The learning method for the artificial neural network model 1100 may include a supervised learning method, in which the model learns to be optimized for solving a problem by inputting correct teacher signals (or labels), and an unsupervised learning method, which does not require teacher signals. According to one embodiment, an electronic device according to one embodiment of the present disclosure (e.g., the electronic device 100 in Figures 1 and 2) can train the artificial neural network model 1100 using video footage of the subject's cardiovascular system and angle information of the video acquisition device relative to the subject.

[0113] According to one embodiment, the electronic device can generate training data for training an artificial neural network model 1100. For example, the electronic device can generate a training dataset that includes video footage of a subject's cardiovascular system and angle information of the video acquisition device relative to the subject. The electronic device can then train an artificial neural network model 1100 to identify at least one type of cardiovascular system contained in the video (or classify the blood vessels in the video) based on the generated training dataset.

[0114] According to one embodiment, the input variables of the artificial neural network model 1100 can include video footage of the subject's cardiovascular system and angle information of the video acquisition device relative to the subject. When the aforementioned input variables are input through the input layer 1120, the output variables output by the output layer 1140 of the artificial neural network model 1100 may be information identifying at least one type of cardiovascular system included in the video (or vascular classification information).

[0115] In this way, the artificial neural network model 1100 can learn to extract the correct output corresponding to a specific input by matching multiple input variables with multiple corresponding output variables in the input layer 1120 and output layer 1140, and by adjusting the synaptic values ​​between nodes included in the input layer 1120, hidden layers 1130_1~1130_n and output layer 1140. Through this learning process, the characteristics hidden in the input variables of the artificial neural network model 1100 can be grasped, and the synaptic values ​​(or weights) between nodes of the artificial neural network model 1100 can be adjusted to reduce the error between the output variable calculated based on the input variable and the target output. Furthermore, the electronic device can learn an algorithm that receives video footage of the subject's cardiovascular system and angle information of the video acquisition device relative to the subject as input, and learn in a way that minimizes the loss with information identifying at least one type of cardiovascular system (or vascular classification information) (i.e., annotation information) included in the video. Using the artificial neural network model 1100 trained in this manner, information identifying at least one type of cardiovascular system contained in the video can be estimated.

[0116] The flowchart and explanation described above are merely illustrative examples and may be implemented differently in some embodiments. For example, in some embodiments, the order of the steps may be changed, some steps may be repeated, some steps may be omitted, or some steps may be added.

[0117] The aforementioned methods may be provided by computer programs stored on computer-readable recording media for execution on a computer. The media may be used to continuously store computer-executable programs, or to temporarily store them for execution or download. Furthermore, the media may be a variety of recording or storage means, often consisting of a single or several hardware components combined, and is not limited to media directly connected to a computer system; it may also be distributed across a network. Examples of media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and media configured to store program instructions, including ROM, RAM, and flash memory. Other examples of media include recording or storage media managed by app stores and other sites and servers that supply or distribute various software.

[0118] The methods, operations, or techniques described herein may be embodied by a variety of means. For example, such techniques may be embodied in hardware, firmware, software, or a combination thereof. A person of ordinary skill will understand that the various exemplary logical blocks, modules, circuits, and algorithmic stages described in conjunction with the disclosure may be embodied in electronic hardware, computer software, or a combination thereof. To clearly illustrate such mutual substitutability between hardware and software, various exemplary components, blocks, modules, circuits, and stages have been generally described above in terms of their functional aspects. Whether such functions are embodied as hardware or software depends on the design requirements imposed on the particular application and the overall system. A person of ordinary skill may embodied the functions described in a variety of ways for their respective specific applications, but such embodiments should not be construed as deviations from the scope of this disclosure.

[0119] In hardware implementations, the processing units used to perform the techniques may be embodied in one or more ASICs, DSPs, digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, microcontrollers, microprocessors, electronic devices, other electronic units designed to perform the functions described herein, computers, or combinations thereof.

[0120] Accordingly, the various exemplary logic blocks, modules, and circuits described in conjunction with this disclosure may be embodied or performed by any combination of general-purpose processors, DSPs, ASICs, FPGAs or other programmable logic devices, discrete gates and transistor logic, discrete hardware components, or any other devices designed to perform the functions described herein. A general-purpose processor may be a microprocessor, but alternatively, the processor may be any conventional processor, controller, microcontroller, or state machine. The processor may also be embodied by a combination of computing devices, such as a DSP and a microprocessor, multiple microprocessors, one or more microprocessors in conjunction with a DSP core, or any other combination of configurations.

[0121] In the embodiment of firmware and / or software, the technique may be embodied in instructions stored on a computer-readable medium such as random access memory (RAM), read-only memory (ROM), non-volatile random access memory (NVRAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable PROM (EEPROM), flash memory, compact disc (CD), or magnetic or marked data storage devices. The instructions may be executable by one or more processors, and may cause the processors(s) to perform specific modes of the functions described herein.

[0122] When embodied in software, the techniques described above may be stored on or transmitted through a computer-readable medium in one or more instructions or codes. A computer-readable medium includes both computer storage and communication media, encompassing any medium that facilitates the transmission of computer programs from one location to another. The storage medium may be any available medium accessible by a computer. As a non-limiting example, such a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to transport or store desired program code in the form of instructions or data structures and is accessible by a computer. Furthermore, any connection may be appropriately made on the computer-readable medium.

[0123] For example, when software is transmitted from a website, server, or other remote source using coaxial cable, fiber optic cable, stranded wire, digital subscriber line (DSL), or wireless technologies such as infrared, radio, and microwave, these are included within the definition of a medium. The terms "disk" and "disc" as used in this application include CDs, laserdiscs, optical discs, DVDs (digital versatile discs), floppy disks, and Blu-ray discs, where "disks" typically reproduce data magnetically and "discs" optically reproduce data using a laser. The aforementioned combinations must also be included within the scope of computer-readable media.

[0124] The software module may reside in RAM memory, flash memory, ROM memory, EPROM memory, EEPROM memory, registers, hard disk, portable disk, CD-ROM, or any other known form of storage medium. An exemplary storage medium may be linked to the processor so that the processor can read information from or record information to the storage medium. Alternatively, the storage medium may be integrated into the processor. The processor and storage medium may reside within an ASIC. The ASIC may reside within a user terminal. Alternatively, the processor and storage medium may exist as separate components in the user terminal.

[0125] Although the embodiments described above are described as utilizing the aspects of the subject matter currently disclosed in one or more standalone computer systems, the disclosure is not limited to and may be embodied in conjunction with any computing environment, such as a network or a distributed computing environment. In fact, aspects of the subject matter may be embodied in multiple processing chips or devices in the disclosure, and storage may be similarly affected across multiple devices. Such devices may include PCs, network servers, and portable devices.

[0126] While this disclosure has been described in relation to some embodiments, various modifications and alterations are possible without departing from the scope of this disclosure as understandable to a person of the ordinary skill in the art to which the invention of this disclosure pertains. Such modifications and alterations should be understood to fall within the scope of the claims appended to this specification. [Explanation of symbols]

[0127] 100:Electronic equipment 210: Memory 220: Processor 230: Communication module 240: Input / Output Interface

Claims

1. In a vascular classification method performed by at least one processor, A step of acquiring images including the cardiovascular system of the subject and angle information of the image acquisition device relative to the subject; A step in which a machine learning model is trained to identify the view corresponding to each of the multiple images of cardiovascular tissue taken as input, based on view information classified by the first and second rotation angle information of the image acquisition device and class information in which the type of blood vessel matched to the view information is specified as one class; A step of identifying a view corresponding to the video based on the machine learning model; and A vascular classification method comprising the step of identifying at least one cardiovascular type contained in the image based on the identified view.

2. The aforementioned angle information is, The first rotation angle information of the image acquisition device centered on the first vertical axis of the subject, and The blood vessel classification method according to claim 1, further comprising second rotation angle information of the image acquisition device centered on a second axis in the left-right direction of the subject.

3. The step of identifying at least one cardiovascular type is, A method for classifying blood vessels according to claim 1, comprising the step of identifying at least one type of cardiovascular vessel included in the video based on the angle information and information on the type of cardiovascular vessel matched with the angle information.

4. The aforementioned class information is The blood vessel classification method according to claim 1, further comprising at least one of the following: a class that enables blood vessel classification for images other than CAG (coronavirus angiography) images among the plurality of images, or a dummy class that includes only the type of blood vessel without the view information.

5. The step of identifying at least one cardiovascular type is, The vascular classification method according to claim 1, further comprising the step of limiting the types of cardiovascular vessels that can be identified in the image based on the identified view.

6. The step of identifying the view corresponding to the aforementioned video is: The step of generating weighted values ​​based on a predetermined number of locations in the aforementioned video; A step of extracting at least one representative frame from the video based on the generated weighted value; The step of inputting at least one representative frame into the machine learning model; and The blood vessel classification method according to claim 1, further comprising the step of identifying a view corresponding to the video based on the inference result for at least one representative frame through the machine learning model.

7. The step of identifying a view corresponding to the video based on the inference result for at least one representative frame is: If the at least one representative frame includes multiple representative frames, the step includes determining the final result through a voting method for the inference results for the multiple representative frames through the machine learning model, The step of determining the final result is, If a valid view is selected based on the voting results for the inference results of the aforementioned multiple representative frames, the final result is to determine a class that includes the type of blood vessel and view information corresponding to the selected valid view; If an invalid view is selected as a result of the aforementioned voting, the inference results for the multiple representative frames are voted on again, excluding the invalid view; if a valid view is selected as a result of the subsequent voting, the final result is determined to be a class containing the blood vessel type and view information corresponding to the selected valid view; If an invalid view is selected as a result of the voting, the inference results for the multiple representative frames are voted on again, excluding the invalid view; if all are invalid views as a result of the voting again, the vote is cast again with a second confidence level; if a valid view is selected as a result of the voting again, the final result is determined to be a class including the type of vessel and view information corresponding to the selected valid view; and The blood vessel classification method according to claim 6, further comprising the step of determining the final result by: if an invalid view is selected as a result of the voting, voting again on the remaining frames excluding the invalid view from the inference results for the plurality of representative frames; if all of the views are invalid as a result of the voting again, voting again with the second confidence level; if the views are invalid or not CAG images as a result of the voting again, voting again with the first confidence level only on the view information of the class containing the blood vessel type from the initial inference result.

8. In electronic devices, memory; and Includes at least one processor connected to the memory and configured to execute at least one computer-readable program contained in the memory, The aforementioned at least one program, The system acquires images including the cardiovascular system of the subject and angle information of the image acquisition device relative to the subject. Based on view information classified by the first and second rotation angle information of the image acquisition device and class information in which the type of blood vessel matched to the view information is specified as a single class, a machine learning model is trained to identify the view corresponding to each of the multiple images in which multiple images of cardiovascular tissue are captured are taken as input. Based on the aforementioned machine learning model, the view corresponding to the video is identified, An electronic device comprising a command for identifying at least one cardiovascular type contained in the image, based on the identified view.

9. The main body has a built-in lifting and lowering drive unit; A lifting unit fixed to the upper end of the aforementioned lifting drive unit and which moves up and down in a first direction; A rotating part is connected to the lifting section at one end so as to be rotatable about a first axis in a second direction perpendicular to the first direction, and has a bent surface formed at the other end; A C-shaped frame portion is provided in a "C" shape and is slidably connected to the aforementioned curved surface; An X-ray generator positioned at one end of the C-shaped frame section; and The electronic device according to claim 8, including an image acquisition device disposed at the other end of the C-shaped frame portion.

10. The aforementioned angle information is, The first rotation angle information of the image acquisition device, which is formed when the rotating part rotates around the first axis, and The electronic device according to claim 9, wherein the C-shaped frame portion is formed when it slides and includes second rotation angle information of the image acquisition device centered on a second axis in a third direction perpendicular to the first and second directions.

11. A computer-readable computer program for performing the method according to any one of claims 1 to 7 on a computer.

Images