Device and method for analyzing medical image

Abstract

A device for analyzing a medical image by using artificial intelligence according to an embodiment: inputs a medical image to a first model to extract a feature vector that a second model uses to perform a first task; inputs the feature vector to the second model to cause the second model to perform the task; provides a first medical result describing a health state based on the medical image on the basis of an output result of the second model; and performs a second task requested by a user of the device for the first medical result, wherein the second task comprises: a first additional task of transferring the first medical result to the outside; and a second additional task of performing secondary analysis including at least one of assisted decision-making and difference analysis on the first medical result.

Description

Device and method for analyzing medical images

[0001] The disclosed embodiments relate to a technology for analyzing medical images using artificial intelligence and supporting users in making decisions based on the analysis results.

[0002] [Cross-reference to related applications]

[0003] This application claims priority to Republic of Korea Provisional Patent Application No. 10-2024-0014325, filed January 30, 2024, the entire contents of which are incorporated herein by reference.

[0004] AI technology is actively utilized in the medical field. For example, doctors use AI to assess a patient's health status. With the recent miniaturization of terminals, this AI technology is now being implemented in personalized devices like smartphones. Consequently, services that allow doctors to analyze patient test results using personalized devices have been developed and are widely used.

[0005] When utilizing these personalized devices, analysis results cannot be used further once provided to the user. Because AI analysis consumes expensive resources, this inefficient use should be avoided. In other words, a solution is needed that allows multiple medical professionals to share the results of a single test, allowing them to be further analyzed and utilized for secondary purposes.

[0006] The disclosed embodiments are intended to analyze medical images using artificial intelligence and support users in making medical-based decisions based on the analysis results.

[0007] In one embodiment, a device for analyzing a medical image using artificial intelligence comprises: one or more processors; and a memory for storing instructions to be executed by the one or more processors, wherein the one or more processors are configured to: input the medical image into a first model, extract a feature vector used by a second model to perform a first task, input the feature vector into the second model, cause the second model to perform the first task, provide a first medical result describing a health condition of the medical image based on an output result of the second model, and perform a second task requested by a user of the device for the first medical result, wherein the second task includes: a first additional task for transmitting the first medical result to the outside; and / or a second additional task for performing a secondary analysis including at least one of auxiliary decision making and difference analysis for the first medical result.

[0008] The medical image may include at least one of an electrocardiogram image and a chest X-ray image acquired by the device.

[0009] The one or more processors may be configured to: when a request for external transmission of the first medical result is received from the user, display information of one or more other users capable of transmission on the screen of the device, and the information of the one or more other users may include at least one of the department and medical institution to which the other user belongs, and transmit the first medical result to a specific user selected by the user among the one or more other users.

[0010] The one or more processors may be configured to: record a transmission record of the first medical result between the user and the specific user in a consultation channel in which the user and the specific user participate, and display at least one of the other user's affiliated department, the affiliated medical institution, the last conversation content of the consultation channel, and the last conversation time on a preview screen of the consultation channel.

[0011] The one or more processors may be configured to display the first medical result on a background screen in the advisory channel.

[0012] The one or more processors are configured to display a list of all medical results including the first medical result and other medical results previously acquired by the device, along with whether the output was successful or not, on the screen of the device, wherein the other medical results may include medical results previously acquired by the device by inputting other medical images into the first model and the second model before inputting the medical images into the first model.

[0013] The one or more processors may be configured to: analyze differences in the first medical result based on a selection order of the user, targeting some medical results selected by the user from the entire medical results, including the first medical result, and visually present the differences based on the selection order.

[0014] The one or more processors may be configured to: display together in the list of the partial medical results a number corresponding to the user's selection order.

[0015] The one or more processors may be configured to provide, as the first medical result, at least one of a heart rhythm type, contractility, biomarker value and distribution, sensitivity, and health-related suggestion corresponding to the medical image.

[0016] The one or more processors are configured to input an original image of the medical image into a third model and process the original image into the medical image, and the third model may be trained to classify the examination type of the input image when an input image is input and process the input image according to a region of interest set corresponding to the examination type.

[0017] The one or more processors may be configured to, when a click event by a user occurs on a first user interface displayed on a screen of the device, call a function connected to the click event, and, through the function, remove the first user interface from the screen and display a second user interface on the screen.

[0018] The second task may include a secondary analysis function accessible to the device through a purchase in an app store where one or more web services are registered, and the secondary analysis function may be executed when a click event occurs on the second user interface.

[0019] A method for analyzing a medical image using artificial intelligence according to one embodiment is a method for analyzing a medical image using artificial intelligence, the method being performed by a device having one or more processors; and a memory for storing instructions to be executed by the one or more processors, the method comprising: inputting the medical image into a first model to extract a feature vector used by a second model to perform a first task; inputting the feature vector into the second model to cause the second model to perform the first task; providing a first medical result describing a health condition of the medical image based on an output result of the second model; performing a second task requested by a user of the device for the first medical result, wherein the performing the second task comprises: performing a first additional task for transmitting the first medical result to an external source; and / or performing a second additional task for performing a secondary analysis including at least one of auxiliary decision making and difference analysis for the first medical result.

[0020] The medical image may include at least one of an electrocardiogram image and a chest X-ray image acquired by the device.

[0021] The step of performing the first additional task may include, when a request for external transmission of the first medical result is received from the user, displaying information of one or more other users capable of transmission on the screen of the device; the information of the one or more other users may include at least one of the department and medical institution to which the other user belongs, and transmitting the first medical result to a specific user selected by the user among the one or more other users.

[0022] The step of performing the first additional task may include: recording a record of the transmission of the first medical result between the user and the specific user in a consultation channel in which the user and the specific user participate; and displaying at least one of the other user's affiliated department, the affiliated medical institution, the last conversation content of the consultation channel, and the last conversation time on a preview screen of the consultation channel.

[0023] The step of performing the first additional task may include the step of displaying the first medical result on the background screen in the advisory channel.

[0024] The step of performing the first additional task includes the step of displaying a list of all medical results including the first medical result and other medical results previously acquired by the device, along with whether or not the output was successful, on the screen of the device, and the other medical results may include medical results previously acquired by the device by inputting other medical images into the first model and the second model before inputting the medical images into the first model.

[0025] The step of performing the second additional task may include a step of analyzing differences in the first medical result based on the user's selection order for some medical results selected by the user from the entire medical results, including the first medical result; and a step of visually presenting the differences based on the selection order.

[0026] The step of performing the second additional task may include the step of displaying a number corresponding to the user's selection order together with the list of some of the medical results.

[0027] The step of providing the first medical result may include a step of providing, as the first medical result, at least one of a type of heart rhythm, contractility, value and distribution of a biomarker, sensitivity, and health-related suggestion corresponding to the medical image.

[0028] The method further includes: inputting an original image of the medical image into a third model, and processing the original image into the medical image, wherein the third model can be trained to classify the examination type of the input image when the input image is input, and process the input image according to a region of interest set corresponding to the examination type.

[0029] The method may include: a step of calling a function connected to a click event when a click event by a user occurs on a first user interface displayed on a screen of the device; and a step of removing the first user interface from the screen and displaying a second user interface on the screen through the function.

[0030] The second task may include additional analytics functionality accessible by the device through a purchase in an app store where one or more web services are registered, wherein the additional analytics functionality may be executed when a click event occurs on the second user interface.

[0031] The disclosed embodiments utilize artificial intelligence to analyze input medical images and provide medical results, thereby providing medical results to users regardless of time and place.

[0032] The disclosed embodiments can prevent waste of resources caused by duplicating the same task by providing a function for transmitting medical results analyzed by one user to another user.

[0033] The disclosed embodiments can provide users with medical-based decision-making suggestions in that, in addition to primary analysis of medical images, they provide secondary analysis including a Clinical Decision Support System (CDSS) and difference analysis.

[0034] FIG. 1 is a block diagram illustrating a device for analyzing a medical image according to one embodiment.

[0035] Figure 2 is an exemplary drawing for explaining the process of inputting a medical image.

[0036] Figures 3a to 3e are exemplary drawings for explaining the sharing request process and result.

[0037] Figures 4a to 4d are exemplary drawings for explaining the display screen of the first medical result.

[0038] Figures 5a to 5d are exemplary drawings for explaining the process and results of a secondary analysis request as a second task.

[0039] Figure 6 is an example diagram illustrating the user interface displayed on the screen of the device.

[0040] FIG. 7 is a flowchart illustrating a method for analyzing a medical image according to one embodiment.

[0041] The terms used in this specification may vary depending on the functions used in the invention, the intentions or customs of the user or operator, etc. Therefore, the definitions should be made based on the contents throughout this specification. The terminology used in the detailed description is only for the purpose of describing one embodiment and should never be limited. Unless clearly used otherwise, the singular form includes the plural form. In this description, the expressions such as "comprises" or "having" are intended to indicate certain components, numbers, steps, operations, elements, parts or combinations thereof, and should not be construed to exclude the presence or possibility of one or more other components, numbers, steps, operations, elements, parts or combinations thereof other than those described.

[0042] Although terms like "first" and "second" are used to describe various components, these components are not limited by these terms. These terms are used merely to distinguish one component from another. Therefore, a "first" component referred to below may also be a "second" component within the technical scope of the present invention.

[0043] Furthermore, the embodiments described herein may have aspects that are entirely hardware, partially hardware and partially software, or entirely software. As used herein, the terms "unit," "apparatus," "module," "device," "server," or "system" refer to hardware, a combination of hardware and software, or a computer-related entity such as software. For example, a unit, apparatus, module, device, server, or system may refer to hardware that constitutes part or all of a platform and / or software such as an application for operating the hardware. As a specific example, a unit, apparatus, module, device, server, or system may be implemented by a processor.

[0044] The embodiments described herein may have aspects that are entirely hardware, partially hardware and partially software, or entirely software.

[0045] FIG. 1 is a block diagram illustrating a device for analyzing a medical image according to one embodiment.

[0046] Referring to FIG. 1, a device for analyzing a medical image includes an extraction unit, a first working unit, a provision unit, and a second working unit.

[0047] The extraction unit inputs a medical image into the first model and extracts a feature vector that the second model uses to perform the first task.

[0048] Medical images are various medical images acquired regardless of the channel, such as black and white and color, and may include, for example, at least one of an electrocardiogram image and a chest X-ray image.

[0049] Meanwhile, medical images can be preprocessed images. For example, medical images can be preprocessed to fit the required size and aspect ratio of the original image, cropped from the original image to a region of interest, pre-screened for medical image-relatedness (e.g., 12-lead images), or images with guaranteed quality values ​​related to noise in the input image.

[0050] Meanwhile, the types of preprocessing described above are merely examples, and all processing steps that can be performed to improve the accuracy of the first model may be included, and are not necessarily limited to the examples described above.

[0051] Feature vectors are medical data that can be extracted from medical images, and may include, for example, electrocardiogram rhythms, physiological and anatomical values, and values ​​related to the risk of disease.

[0052] The first model is a vision-based deep learning model trained to extract feature vectors from medical images, and may be, for example, a model designed based on at least one of Modified CNNs, EfficientNet, ResNext, and Transformer.

[0053] The first task inputs a feature vector into the second model, and causes the second model to perform the task.

[0054] Here, the second model may be a model trained to perform various tasks such as regression, classification, or event prediction based on the input feature vector.

[0055] Specifically, the second model may be a model trained to predict at least one of the heart rhythm type, contractility, biomarker value and distribution, sensitivity, and health-related suggestions based on the feature vector output by the first model.

[0056] The provider provides a first medical result describing the health status of the medical image based on the output of the second model. The first medical result is a result report organized for easy interpretation by the user, and may include at least one of rhythm type, contractility, biomarker values ​​and distribution, sensitivity, and health-related suggestions.

[0057] The second task performs the second task requested by the user of the device for the first medical result.

[0058] The second task may include a first additional task that communicates the first medical result externally.

[0059] Specifically, the first additional task may include processing a request from a user to communicate the first medical result externally (e.g., to another user).

[0060] At this time, the second task unit displays information of other users on the screen of the device, for example, as a list, and when a user who has checked the list selects a specific user, the second task unit can transmit the first medical result to the selected specific user.

[0061] Meanwhile, the list can include other users' affiliated medical institutions and departments. This provides users with helpful information to help them communicate their initial medical results to the appropriate medical professional.

[0062] Here, the second task can record the transmission record in a consultation channel where the user and other users participate. The second task can display the first medical result on the background screen of the chat room. Furthermore, the second task can display at least one of the other user's affiliated department, affiliated medical institution, the last conversation in the consultation channel, and the time of the last conversation on the preview screen of the consultation channel.

[0063] The second task may include a second additional task that performs a secondary analysis including at least one of auxiliary decision making and difference analysis on the first medical outcome.

[0064] The second task force can provide supplementary decision-making through further analysis of the first medical outcome.

[0065] Here, assisted decision making refers to information that helps support clinical decisions based on the CDSS and may include useful information during patient diagnosis and treatment. Specifically, assisted decision making may include recommendations, warnings, and guidelines appropriate to the user's condition.

[0066] The second task can compare the first medical outcome with other first medical outcomes and analyze the differences in the first medical outcome.

[0067] Here, the first other medical result may mean a past medical result for the same user previously acquired by the device, which is information generated by inputting the first model and the second model into another medical image before the first medical result is generated.

[0068] The second work unit can display a list of first medical results in chronological order on the screen of the device based on the work history.

[0069] The second task unit can display the entire medical result list, including the first medical result and other medical results previously acquired by the device, on the device's screen, indicating whether the result was successfully output. The second task unit can also display numbers corresponding to the user's selection order along with the list of some medical results.

[0070] The second task unit can analyze differences in the first medical result, based on the user's selection order, for selected medical results, including the first medical result, from the overall medical results. The second task unit can visually present the differences based on the selection order.

[0071] Figure 2 is an exemplary drawing for explaining the process of inputting a medical image.

[0072] Referring to FIG. 2, the screen of the device includes a medical image. Here, the medical image is an image provided by the user or previously stored in the device and selected by the user as an analysis target, such as an electrocardiogram or a chest X-ray image.

[0073] Medical images may be original images selected by the user or images processed by automatically detecting regions of interest (e.g., regions associated with an electrocardiogram or parts associated with a chest X-ray) from the original images.

[0074] Specifically, the device can automatically detect a region of interest in an input image using a third model. For example, if the device receives a raw medical image, the device can first input the raw image into a third model to obtain a processed medical image that retains only the region of interest from the original image.

[0075] At this time, the third model may be a model trained to classify the examination type of the input image when an input image is input, and process the input image according to the region of interest set in response to the examination type. For example, the third model may classify the examination type of the input image, such as whether the input image is a chest X-ray image or an electrocardiogram image showing the waveform of the lead. Here, the third model may also classify the examination type of the input image by learning the pattern of medical data appearing in the input image. Thereafter, the third model may automatically extract the region of interest set according to the examination type and process the region of interest into a medical image. As a specific example, the third model may automatically detect the region of interest by extracting an object in the shape of a box from the input image, such as a CNN-based detection model, YOLO, or Faster R-CNN.

[0076] Meanwhile, the area of ​​interest can be detected automatically, but can also be manually set based on some points (2 or 4 points) selected by the user.

[0077] Figures 3a to 3d are exemplary drawings for explaining the sharing request process and result.

[0078] Referring to FIG. 3a, the screen of the device shows a list of information about other users who can transmit the first medical result.

[0079] Users can be identified by their activity status through an activation button. For example, if the activation button corresponding to a user is activated, the user is identified as being able to respond to a consultation in the consultation channel.

[0080] Users whose activation button is activated may have their personal information displayed together in the list, for example, whether the user's affiliated institution is 'Internal Medicine' or 'Emergency Medicine', or whether the affiliated hospital is 'Bundang Seoul National University Hospital', etc. may be displayed together in the list.

[0081] Referring to Figure 3b, items that can be entered together with a user's transmission request are illustrated.

[0082] For example, a user may request that the first medical result be delivered to a specific user, along with additional information of the user's choosing.

[0083] For example, a user may request that at least one of the following be transmitted to another user along with the first medical result: the time the first medical result was taken, the patient's age, gender, type of pain, and other comments.

[0084] Referring to Figure 3c, a preview screen of the advisory channel is shown.

[0085] At this time, the preview of the advisory channel is the preview screen of the advisory channel displayed after the transmission request in Figure 3b is completed. The preview screen of the advisory channel may display information about the time the transmission request was completed and other users who received the first medical result (e.g., consultants who can respond to the consultation).

[0086] Referring to Figure 3d, the screen of the consultation channel is illustrated. The screen of the consultation channel may display the first medical result exchanged between a user (e.g., a client requesting advice) and another user (e.g., a consultant responding to the advice) on the background of the consultation channel.

[0087] From this, a user (e.g., a client requesting advice) can consult advice while simultaneously checking the first medical result shared by another user (e.g., a consultant who responded to the advice).

[0088] Referring to FIG. 3e, a list of medical results is displayed on the screen of the device.

[0089] The screen may display a list of medical results, along with numbers corresponding to the order in which the user requested and selected the medical results. The screen may also indicate whether the medical results were successfully received. If the medical results were not received, a symbol, such as an "x," may be displayed on the list to indicate whether the results were received.

[0090] Figures 4a to 4d are exemplary drawings for explaining the display screen of the first medical result.

[0091] Referring to FIG. 4A, the device can display the first medical result on its screen while simultaneously displaying other advertising information on the screen. In this case, the device can display an advertisement related to the first medical result on the screen along with the first medical result.

[0092] The device provides a page-by-page analysis of the first medical result. Page B, the first page, displays the medical image selected by the user for analysis, as shown.

[0093] For example, the device may display, as a first medical result, the type of rhythm (e.g., sinus rhythm) and contractility (e.g., left ventricular ejection fraction, left ventricular, right ventricular, and left atrium strain (GLS)) corresponding to the medical image. Furthermore, the device may display biomarkers and their values ​​corresponding to the medical image on the screen together.

[0094] Meanwhile, the device can display medical images full-screen when touched for a long time. At this time, the medical image can be appropriately displayed (horizontal or vertical) based on the screen size.

[0095] Referring to FIG. 4b, the device can divide the expected sensitivity according to a specific threshold in the entire biomarker section into four grades as shown on the second page, S page, and display the expected sensitivity for each grade, the optimal threshold using Youden's Index, etc., and the sensitivity specificity at that time as a background. Referring to FIG. 4c, the device can visually display the distribution of biomarkers in a specific population group as a background as shown on the third page, Q page.

[0096] The distribution of biomarkers corresponding to medical images can be measured by measuring biomarkers in a specific patient population. Based on this distribution, interquartile intervals for the biomarkers can be calculated, and each interval can be displayed as a graph with a different color in the background.

[0097] Referring to FIG. 4d, the device can display the values ​​of biomarkers in detail on the screen and display preset medical suggestions based on the combination of values ​​of the biomarkers.

[0098] Figures 5a to 5d are exemplary drawings for explaining the process and results of a secondary analysis request as a second task.

[0099] Referring to Figure 5a, the device can predict auxiliary decision-making based on additional information added to the first medical result at the user's request. The additional information may include clinical information entered by the user.

[0100] Referring to FIG. 5b, it can be seen that the screen of the device displays the first medical result along with the auxiliary decision.

[0101] Referring to Figure 5c, the device, at the user's request, displays a trend of the first medical result. At this time, the device compares multiple first medical results and visually presents the differences in the first medical results.

[0102] Referring to FIG. 5D, the device can compare a plurality of first medical results obtained in advance to visually provide differences in the first medical results. Here, the numbers located on the x-axis of the graph indicate the analysis order selected by the user as the analysis target. For example, 1 is a symbol assigned to identify the medical result that the user first sets as the analysis target, 2 is a symbol assigned to identify the medical result that the user second sets as the analysis target, and 3 is a symbol assigned to identify the medical result that the user third sets as the analysis target, wherein it is preferable that one of the medical results includes the first medical result.

[0103] Figure 6 is an example diagram illustrating the user interface displayed on the screen of the device.

[0104] As illustrated in FIG. 6, when a first user interface is displayed on the screen of the device and a click event for the first user interface occurs by the user, the device can call a function connected to the click event.

[0105] The device can, through a function, remove the first user interface from the screen and display a second user interface on the screen. The first user interface may be a quick menu in the form of a drop-down menu that lists functions that the device can process. The second user interface may include functions that the device can process, such as a "Purchase Menu," a "Favorites Menu," or a menu for additional analysis functions that the device can process by purchasing them through the App Store.

[0106] FIG. 7 is a flowchart illustrating a method for analyzing a medical image according to one embodiment.

[0107] Referring to FIG. 7, the method of FIG. 6 can be performed by the device of FIG. 1.

[0108] First, a device for analyzing a medical image according to one embodiment inputs a medical image into a first model and extracts a feature vector used by the second model to perform the first task.

[0109] Thereafter, the device for analyzing a medical image according to one embodiment inputs a feature vector into a second model, thereby causing the second model to perform the first task.

[0110] Thereafter, the device for analyzing a medical image according to one embodiment provides a first medical result describing a health condition of the medical image based on the output result of the second model.

[0111] Thereafter, the device for analyzing a medical image according to one embodiment performs a second task requested by the user of the device for the first medical result.

[0112] Here, the second task includes a first additional task that transmits the first medical result to an external source; and a second additional task that performs a secondary analysis including at least one of auxiliary decision making and difference analysis on the first medical result.

[0113] Meanwhile, although the method of FIG. 7 is described as being divided into multiple steps, at least some of the steps may be performed in a different order, combined with other steps and performed together, omitted, divided into detailed steps and performed, or one or more steps not shown may be added and performed.

[0114] While representative embodiments of the present invention have been described in detail above, those skilled in the art will appreciate that various modifications to the above-described embodiments are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the claims set forth below but also by equivalents thereof.

[0115] A device and method for analyzing a medical image according to one embodiment analyzes a medical image using an artificial intelligence model, and can be used in the digital medical industry.

Claims

1. One or more processors; and A device for analyzing medical images using artificial intelligence, comprising a memory for storing commands executed by one or more processors, One or more of the above processors: By inputting the above medical image into the first model, the feature vector used by the second model to perform the first task is extracted, By inputting the feature vector into the second model, the second model performs the first task, Provide a first medical result describing the health status of the medical image based on the output result of the second model, configured to perform a second operation requested by the user of the device for the first medical result; The second task above is: A first additional operation for transmitting the first medical result to an external party; and / or A device for analyzing medical images using artificial intelligence, comprising a second additional task of performing secondary analysis including at least one of auxiliary decision making and difference analysis for the first medical result.

2. In paragraph 1, A device for analyzing medical images using artificial intelligence, wherein the medical image includes at least one of an electrocardiogram image and a chest X-ray image acquired by the device.

3. In paragraph 1, One or more of the above processors: When a request for external transmission of the first medical result is received from the user, information of one or more other users for which transmission is possible is displayed on the screen of the device, The information of one or more other users includes at least one of the other users' affiliated departments and affiliated medical institutions, A device for analyzing medical images using artificial intelligence, configured to transmit the first medical result to a specific user selected by the user among the one or more other users.

4. In paragraph 3, One or more of the above processors: Record the transmission record of the first medical result between the user and the specific user in the consultation channel in which the user and the specific user participate, A device for analyzing medical images using artificial intelligence, configured to display at least one of the other user's affiliated department, the affiliated medical institution, the last conversation content of the advisory channel, and the last conversation time on the preview screen of the advisory channel.

5. In paragraph 4, One or more of the above processors, A device for analyzing medical images using artificial intelligence, configured to display the first medical result on the background screen in the above-mentioned advisory channel.

6. In paragraph 1, One or more of the above processors: The device is configured to display a list of all medical results including the first medical result and other medical results previously acquired by the device, along with whether the output was successful or not, on the screen of the device. A device for analyzing medical images using artificial intelligence, wherein the other medical results include medical results previously acquired by the device by inputting other medical images into the first model and the second model before inputting the medical images into the first model.

7. In paragraph 6, One or more of the above processors: Among the above overall medical results, some medical results selected by the user, including the first medical result, are analyzed based on the user's selection order for the difference in the first medical result. A device for analyzing medical images using artificial intelligence, configured to visually provide the above differences based on the above selection order.

8. In paragraph 7, One or more of the above processors: A device for analyzing medical images using artificial intelligence, configured to display numbers corresponding to the user's selection order together in a list of some of the medical results.

9. In paragraph 1, One or more of the above processors: A device for analyzing a medical image using artificial intelligence, configured to provide at least one of the following: a type of heart rhythm, contractility, value and distribution of biomarkers, sensitivity, and health-related suggestions corresponding to the medical image, as the first medical result.

10. In paragraph 1, One or more of the above processors, The original image of the above medical image is input into a third model, and the original image is configured to be processed into the above medical image, The third model is a device for analyzing medical images using artificial intelligence, which is trained to classify the examination type of the input image when an input image is input and process the input image according to a region of interest set corresponding to the examination type.

11. In paragraph 1, One or more of the above processors, When a click event by a user occurs on the first user interface displayed on the screen of the above device, a function connected to the click event is called, A device for analyzing medical images using artificial intelligence, configured to remove a first user interface from the screen and display a second user interface on the screen through the above function.

12. In paragraph 11, The second task includes a secondary analysis function accessible by the device through a purchase in an app store where one or more web services are registered, The above secondary analysis function is a device for analyzing medical images using artificial intelligence, which is executed when a click event occurs on the above second user interface.

13. One or more processors; and A method for analyzing medical images using artificial intelligence, the method being performed by a device having a memory for storing instructions executed by one or more processors, The above method is: A step of inputting the above medical image into a first model and extracting a feature vector used by the second model to perform the first task; A step of inputting the feature vector into the second model to cause the second model to perform the first task; A step of providing a first medical result describing a health condition of the medical image based on the output result of the second model; A step of performing a second operation requested by a user of the device for the first medical result is included, The step of performing the above second task is: A step of performing a first additional task of transmitting the first medical result to an external party; and / or A method for analyzing medical images using artificial intelligence, comprising the step of performing a second additional operation that includes at least one of auxiliary decision making and difference analysis for the first medical result.

14. In paragraph 13, A method for analyzing a medical image using artificial intelligence, wherein the medical image includes at least one of an electrocardiogram image and a chest X-ray image acquired by the device.

15. In paragraph 13, The step of performing the above first additional task is: When a request for external transmission of the first medical result is received from the user, a step of displaying information of one or more other users capable of transmission on the screen of the device; The information of one or more other users includes at least one of the other users' affiliated departments and affiliated medical institutions, A method for analyzing medical images using artificial intelligence, comprising the step of transmitting the first medical result to a specific user selected by the user among the one or more other users.

16. In paragraph 15, The step of performing the above first additional task is: A step of recording a record of the transmission of the first medical result between the user and the specific user in an advisory channel in which the user and the specific user participate; and A method for analyzing medical images using artificial intelligence, comprising the step of displaying at least one of the other user's affiliated department, the affiliated medical institution, the last conversation content of the advisory channel, and the last conversation time on the preview screen of the advisory channel.

17. In paragraph 16, The step of performing the above first additional task is: A method for analyzing medical images using artificial intelligence, comprising the step of displaying the first medical result on a background screen in the above-mentioned advisory channel.

18. In paragraph 13, The step of performing the above first additional task is: A step of displaying a list of all medical results including the first medical result and other medical results previously acquired by the device on the screen of the device, along with whether or not the output was successful, A method for analyzing a medical image using artificial intelligence, wherein the other medical results include medical results previously acquired by the device by inputting other medical images into the first model and the second model before inputting the medical image into the first model.

19. In paragraph 18, The step of performing the above second additional task is: A step of analyzing the difference for the first medical result based on the selection order of the user, targeting some medical results selected by the user from the entire medical results, including the first medical result; and A method for analyzing medical images using artificial intelligence, comprising the step of visually providing the above difference based on the above selection order.

20. In paragraph 19, The step of performing the above second additional task is: A method for analyzing medical images using artificial intelligence, comprising the step of displaying numbers corresponding to the user's selection order together in a list of some of the medical results.

21. In paragraph 13, The first step in providing medical results is: A method for analyzing a medical image using artificial intelligence, comprising the step of providing at least one of a heart rhythm type, contractility, value and distribution of a biomarker, sensitivity, and health-related suggestion corresponding to the medical image as the first medical result.

22. In paragraph 13, The above method is: Further comprising a step of inputting the original image of the above medical image into a third model and processing the original image into the above medical image, The third model is a method for analyzing medical images using artificial intelligence, wherein when an input image is input, the third model is trained to classify the examination type of the input image and process the input image according to a region of interest set corresponding to the examination type.

23. In paragraph 13, The above method is: When a click event by a user occurs on a first user interface displayed on the screen of the device, a step of calling a function connected to the click event; and A method for analyzing medical images using artificial intelligence, comprising the step of removing a first user interface from the screen and displaying a second user interface on the screen through the above function.

24. In paragraph 23, The second task includes additional analytics capabilities accessible by the device through purchases from an app store where one or more web services are registered, A method for analyzing medical images using artificial intelligence, wherein the above additional analysis function is executed when a click event occurs on the second user interface.

Images