Apparatus, method, system, and program for predicting the occurrence time of acute cerebral infarction

Abstract

The present disclosure includes a communication unit that communicates with an imaging device that captures images of patients with acute cerebral infarction, and a processor that controls operations related to predicting the time of occurrence of acute cerebral infarction, wherein the processor receives images of patients with acute cerebral infarction from the imaging device via the communication unit, classifies the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, acquires the first image and the second image, aligns the first image and the second image with an MNI area corresponding to an activated area of ​​the brain, detects an infarct area in the second image, and outputs a probability prediction value for the time of occurrence of acute cerebral infarction based on the first image, the second image, and the infarct area.

Description

[Technical Field]

[0001] The present disclosure relates to an apparatus, method, system, and program for predicting the onset time of acute cerebral infarction. [Background technology]

[0002] Thrombolytic therapy, which opens cerebral blood vessels blocked by thrombus or embolus and prevents damage to brain cells, is widely known as a method for treating patients with acute cerebral infarction. Thrombolytic therapy must be performed promptly before brain cells are damaged, so it must be performed early from the time of the onset of cerebral infarction.

[0003] However, the majority of patients who visit hospitals without knowing when their acute cerebral infarction occurred have not been diagnosed. In other words, they are discovered by others after collapsing and taken to the hospital, so it is not possible to determine when the acute cerebral infarction occurred before they were discovered. In such cases, there is currently a shortage of specialists who can determine whether a patient is suitable for thrombolytic therapy based solely on medical image interpretation. Furthermore, emergency rooms where acute cerebral infarction patients visit the emergency department are often staffed by medical staff with insufficient experience.

[0004] Therefore, it is necessary to provide information that allows medical staff to safely and quickly determine when acute cerebral infarction occurs. Summary of the Invention [Problem to be solved by the invention]

[0005] The present disclosure has been made in consideration of the above circumstances, and its purpose is to provide information that enables medical staff to safely and quickly determine the time point at which acute cerebral infarction occurs.

[0006] The problems that the present disclosure aims to solve are not limited to those mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below. [Means for solving the problem]

[0007] An acute cerebral infarction occurrence time prediction device according to one aspect of the present disclosure for achieving the above-mentioned technical problem includes a communication unit that communicates with an imaging device that takes images of a patient with acute cerebral infarction, and a processor that controls operations related to predicting the time of occurrence of acute cerebral infarction, wherein the processor receives images of the patient with acute cerebral infarction from the imaging device via the communication unit, classifies the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, acquires the first image and the second image, aligns the first image and the second image with an MNI area corresponding to an activated area of ​​the brain, detects an infarction area in the second image, and outputs a probability prediction value for the time of occurrence of the acute cerebral infarction based on the first image, the second image, and the infarction area.

[0008] The processor may also be characterized by performing machine learning-based learning based on the first image, the second image, and the infarct area, and outputting a probability prediction value for the time of occurrence of acute cerebral infarction that has been learned and analyzed based on the machine learning.

[0009] Furthermore, the processor may be characterized by standardizing the intensity of the second image, performing machine learning-based learning based on the intensity-standardized second image, and detecting the infarct area in the second image that has been learned and analyzed based on the machine learning.

[0010] The processor can also be characterized by labeling the area of ​​the infarct region, performing machine learning-based learning based on the first image, the second image, and the area of ​​the infarct region, and outputting a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

[0011] Furthermore, the processor may further generate reference data for the machine learning platform based on the first image and the second image.

[0012] In addition, a method for predicting the time of occurrence of acute cerebral infarction performed by a prediction device according to another aspect of the present disclosure may include the steps of receiving images of a patient with acute cerebral infarction from an imaging device via a communication unit of the prediction device, classifying the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image via a processor of the prediction device, acquiring the first image and the second image via the processor, aligning the first image and the second image with an MNI region corresponding to an activated region of the brain via the processor, detecting an infarct region in the second image via the processor, and outputting a probability prediction value for the time of occurrence of the acute cerebral infarction based on the first image, the second image, and the infarct region via the processor.

[0013] Furthermore, a system for predicting the time of occurrence of acute cerebral infarction according to another aspect of the present disclosure includes an imaging device that acquires images of a patient with acute cerebral infarction, and an acute cerebral infarction time prediction device that communicates with the imaging device, wherein the prediction device receives the images of the patient with acute cerebral infarction from the imaging device, classifies the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, acquires the first image and the second image, aligns the first image and the second image with an MNI area corresponding to an activated area of ​​the brain, detects an infarction area in the second image, and outputs a probability prediction value for the time of occurrence of the acute cerebral infarction based on the first image, the second image, and the infarction area.

[0014] In addition, a computer program can be provided that is stored in a computer-readable storage medium and causes a computer (hardware) to execute the method for predicting the golden hour at the time of occurrence of acute cerebral infarction.

[0015] In addition, a computer-readable recording medium can be provided that records a computer program that causes a computer to execute a method for realizing the present disclosure. [Effects of the Invention]

[0016] According to the means for solving the above-mentioned problems of the present disclosure, it is possible to obtain the effect of providing information that enables medical staff to safely and quickly determine the time point at which acute cerebral infarction occurs.

[0017] The effects of the present disclosure are not limited to those mentioned above, and other effects not mentioned above will be clearly understood by those skilled in the art from the description below. [Brief explanation of the drawings]

[0018] [Figure 1] 1 is a diagram showing an example of a system for predicting the occurrence time of acute cerebral infarction according to the present disclosure. [Figure 2] The configuration of the prediction device in FIG. [Figure 3] FIG. 3 is a diagram showing an example of a process for outputting a probability prediction value for the time point of occurrence of acute cerebral infarction via the processor of FIG. 2. [Figure 4] FIG. 3 is a diagram showing an example of a process for outputting a probability prediction value for the time point of occurrence of acute cerebral infarction via the processor of FIG. 2. [Figure 5] 1 is a flowchart showing a method for predicting the onset time of acute cerebral infarction according to the present disclosure. [Figure 6] FIG. 6 is a diagram showing an example of a process of outputting a probability prediction value for the time point of occurrence of acute cerebral infarction at the output stage of FIG. 5. DETAILED DESCRIPTION OF THE INVENTION

[0019] The same reference numerals refer to the same components throughout this disclosure. This disclosure does not describe all elements of each embodiment, and general content in the technical field to which this disclosure pertains or overlapping content in the embodiments will be omitted. The terms "unit, module, component, block" used in this specification can be realized by software or hardware, and depending on the embodiment, multiple "units, modules, components, blocks" may be realized as a single component, or one "unit, module, component, block" may include multiple components.

[0020] Throughout this specification, when a part is said to be "connected" to another part, this includes not only direct connection but also indirect connection, including connection via a wireless communication network.

[0021] Furthermore, when a part is described as "comprising" a certain element, this does not mean that it excludes other elements, but that it may further include other elements, unless otherwise specified.

[0022] Throughout this specification, when an element is said to be "on" another element, this includes not only when the element is in contact with the other element, but also when there is another element between the two elements.

[0023] The terms "first," "second," etc. are used to distinguish one component from another, and the components are not limited to the terms described above.

[0024] The singular expression includes the plural expression unless the context clearly indicates otherwise.

[0025] The identification numbers used in each step are for convenience of explanation, and do not dictate the order of the steps; the steps may be performed in a different order than specified unless the context clearly dictates a particular order.

[0026] The working principle and embodiments of the present disclosure will be described below with reference to the accompanying drawings.

[0027] In this specification, the prediction device according to the present disclosure includes various devices capable of performing calculations and providing results to a user. For example, the control unit according to the present disclosure may include all or any one of a computer, a server device, and a portable terminal.

[0028] Here, the computer may include, for example, a notebook computer, a desktop computer, a laptop computer, a tablet PC, a slate PC, etc. equipped with a web browser.

[0029] The server device is a server that communicates with external devices and processes information, and may include an application server, a computing server, a database server, a file server, a mail server, a proxy server, a web server, and the like.

[0030] The portable terminal is, for example, a wireless communication device that ensures portability and mobility, and may include all kinds of handheld-based wireless communication devices such as PCS (Personal Communication System), GSM (Global System for Mobile communications), PDC (Personal Digital Cellular), PHS (Personal Handyphone System), PDA (Personal Digital Assistant), IMT (International Mobile Telecommunication)-2000, CDMA (Code Division Multiple Access)-2000, W-CDMA (W-Code Division Multiple Access), WiBro (Wireless Broadband Internet) terminals, smartphones, etc., as well as wearable devices such as watches, rings, bracelets, anklets, necklaces, glasses, contact lenses, or head-mounted devices (HMDs).

[0031] The system for predicting the occurrence time of acute cerebral infarction according to the present disclosure can be provided to receive images of a patient with acute cerebral infarction from an imaging device, classify the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, acquire the first image and the second image, align the first image and the second image with MNI regions corresponding to activated regions of the brain, detect an infarct region in the second image, and output a probability prediction value for the occurrence time of acute cerebral infarction based on the first image, the second image, and the infarct region.

[0032] The system for predicting the time when acute cerebral infarction occurs according to the present disclosure can provide information that enables medical staff to safely and quickly determine the time when acute cerebral infarction occurs.

[0033] The acute cerebral infarction onset prediction system according to the present disclosure will be described in detail below.

[0034] Fig. 1 is a diagram showing an example of a system for predicting the time of onset of acute cerebral infarction according to the present disclosure, and Fig. 2 shows the configuration of the prediction device of Fig. 1.

[0035] Figures 3 and 4 are diagrams showing an example of a process for outputting a probability prediction value for the time point of occurrence of acute cerebral infarction via the processor of Figure 2. Figure 5 is a diagram showing another example of the system for predicting the time point of occurrence of acute cerebral infarction according to the present disclosure.

[0036] Referring to FIGS. 1-5, a system 1000 may include an image capture device 100 and a prediction device 200 .

[0037] The imaging device 100 can acquire images of patients with acute cerebral infarction and transmit them to the prediction device 200. At this time, the imaging device 100 can acquire magnetic resonance imaging (MRI) images of patients with acute cerebral infarction and transmit them to the prediction device 200. The prediction device 200 can predict the time of occurrence of acute cerebral infarction for patients with acute cerebral infarction. Here, the prediction device 200 can output a probability prediction value for the time of occurrence of acute cerebral infarction. At this time, the prediction device 200 can include a communication unit 210 and a control unit 220.

[0038] The communication unit 210 can communicate with the image capturing device 100 that captures images of acute cerebral infarction patients. The communication unit 210 can receive images of acute cerebral infarction patients acquired from the image capturing device 100. The communication unit 210 can include at least one of a wired communication module and a wireless communication module.

[0039] The wired communication module may include various wired communication modules such as a local area network (LAN) module, a wide area network (WAN) module, or a value added network (VAN) module, as well as various cable communication modules such as a universal serial bus (USB), a high definition multimedia interface (HDMI), a digital visual interface (DVI), a recommended standard 232 (RS-232), a power line communication module, or a plain old telephone service (POTS).

[0040] The wireless communication module may include a Wifi (registered trademark) module, a Wibro (Wireless Broadband) module, as well as a wireless communication module that supports various wireless communication methods such as GSM (global System for Mobile Communication), CDMA (Code Division Multiple Access), WCDMA (Wideband Code Division Multiple Access), UMTS (universal mobile telecommunications system), TDMA (Time Division Multiple Access), LTE (Long Term Evolution), 4G, 5G, and 6G.

[0041] The control unit 220 may include a memory 221 and a processor 222 .

[0042] The memory 221 can store data for an algorithm for controlling the operation of the components in the device or a program that implements the algorithm. The processor 222 can perform the above-mentioned operations using the data stored in the memory 221. Here, the memory 221 and the processor 222 can be implemented as separate chips. Alternatively, the memory 221 and the processor 222 can be implemented as a single chip.

[0043] The memory 221 can store data supporting various functions of the device, programs for operating the components of the device, and input / output data, as well as a number of application programs (or applications) that run on the device, and data and commands for operating the device. At least some of these application programs can be downloaded from an external server via wireless communication.

[0044] Such memory 221 may include at least one type of storage medium among flash memory type, hard disk type, solid state disk type (SSD type), silicon disk drive type (SDD type), multimedia card micro type, card type memory (e.g., SD or XD memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, and optical disk.

[0045] The memory 221 can store images (P1, P2, P3, . . . ) of patients with acute cerebral infarction. The memory 221 can store data related to prediction of the time of occurrence of acute cerebral infarction.

[0046] The processor 222 can control operations related to predicting the time of occurrence of acute cerebral infarction. The processor 222 can receive images (P1, P2, P3, . . .) of an acute cerebral infarction patient from the imaging device 100 via the communication unit 210, classify the images (P1, P2, P3, . . .) into a first image corresponding to a fluid-attenuated inversion recovery (FLAIR) image and a second image corresponding to a diffusion-weighted image (DWI), and acquire the first and second images. In this case, the first and second images may be of an acute cerebral infarction patient whose time of occurrence of acute cerebral infarction has not been confirmed.

[0047] The processor 222 can detect an infarct area in the second image by aligning the first image and the second image with Montreal Neurological Institute (MNI) regions corresponding to activated areas of the brain. Here, the MNI regions can include MNI coordinate values ​​output as result values ​​for activated areas of the brain when analyzing the brain image. In this case, the infarct area can be an area where normal blood supply to growing tissue is impaired due to narrowing or blockage of blood vessels caused by thrombus, embolism, etc.

[0048] The memory 221 can store a probability prediction value for the time point of occurrence of acute cerebral infarction that has been learned and output based on machine learning. Referring to FIG. 4 , the processor 222 can learn input values, such as first image data ID1 corresponding to a water-suppressed image, second image data ID2 corresponding to a diffusion-weighted image, infarction region data ID3, and clinical data ID4 of acute cerebral infarction, based on the machine learning model AIM, and output a recommended result value of the probability prediction value OD for the time point of occurrence of acute cerebral infarction. In this case, the processor 222 can generate machine learning-based reference data based on the first image and the second image. That is, the processor 222 can generate machine learning model AIM-based reference data based on the first image data ID1 corresponding to a water-suppressed image and the second image data ID2 corresponding to a diffusion-weighted image.

[0049] The machine learning model AIM may be constructed to learn through correlations between various first image data ID1, second image data ID2, infarction region data ID3, and acute cerebral infarction clinical data ID4 included in the input data. The machine learning model AIM may be constructed as a learning dataset using a learning algorithm and subjected to reinforcement learning using the various first image data ID1, second image data ID2, infarction region data ID3, and acute cerebral infarction clinical data ID4. In this case, the memory 221 may store a probability prediction value OD for the time point of acute cerebral infarction learned and analyzed based on the machine learning model AIM. Here, the acute cerebral infarction clinical data ID4 may include at least one of various morphological findings, the patient's age at diagnosis, and the location of acute cerebral infarction.

[0050] The input unit 230 can receive first image data corresponding to a water-suppressed image of a patient with acute cerebral infarction and second image data corresponding to a diffusion-weighted image of the patient with acute cerebral infarction. The input unit 230 can also receive infarction region data of the patient with acute cerebral infarction and clinical data of the patient with acute cerebral infarction. The clinical data of acute cerebral infarction can include at least one of various morphological findings, the patient's age at diagnosis, and the location of the acute cerebral infarction. For example, the input unit 230 can be any input means capable of scanning and inputting the first image data corresponding to a water-suppressed image, the second image data corresponding to a diffusion-weighted image, and infarction region data, and inputting clinical data of acute cerebral infarction.

[0051] The processor 222 can output a probability prediction value for the time point of occurrence of acute cerebral infarction based on the first image data, the second image data, and the infarct region data. The processor 222 can also output a probability prediction value for the time point of occurrence of acute cerebral infarction based on the first image data, the second image data, the infarct region data, and clinical data on acute cerebral infarction. The processor 222 can further label the area of ​​the infarct region, perform machine learning-based learning based on the first image data, the second image data, and the area data of the infarct region, and further output a probability prediction value for the time point of occurrence of acute cerebral infarction learned and analyzed based on machine learning. The processor 222 can further label the area of ​​the infarct region by position, perform machine learning-based learning based on the first image data, the second image data, and the area data of the infarct region by position, and further output a probability prediction value for the time point of occurrence of acute cerebral infarction learned and analyzed based on machine learning. Furthermore, the processor 222 can further label the area of ​​the metastasis state by location of the infarct region, perform machine learning-based learning based on the first image data, the second image data, and the area data of the metastasis state by location of the infarct region, and further output a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on machine learning.

[0052] In this case, the processor 222 may standardize the intensity of the second image, perform machine learning-based learning based on the standardized second image, and detect the infarct area in the second image learned and analyzed based on machine learning. When the second image corresponding to such a standardized diffusion-weighted image is learned and analyzed based on the machine learning model AIM via the processor 222, the readability is increased and the image quality can be improved.

[0053] The processor 222 can output the result of the probability prediction value for the recommended occurrence time point of acute cerebral infarction corresponding to the input information of the input unit 230. The display unit 240 can display the probability prediction value for the occurrence time point of acute cerebral infarction output by the processor 222. For example, the display unit 240 can display the probability prediction values ​​belonging to each interval of the occurrence time point of acute cerebral infarction: 0 to 3 hours, 3 to 4.5 hours, 4.5 to 6 hours, 6 to 12 hours, and over 12 hours.

[0054] Fig. 5 is a flowchart showing a method for predicting the occurrence time of acute cerebral infarction according to the present disclosure. Fig. 6 is a diagram showing an example of a process for outputting a probability prediction value for the occurrence time of acute cerebral infarction in the output stage of Fig. 5.

[0055] Referring to Figures 5 and 6, the method for predicting the time of occurrence of acute cerebral infarction may include a receiving step (S510), a classification step (S520), an acquisition step (S530), a matching step (S540), a detection step (S550), and an output step (S560).

[0056] In the receiving step, images (P1, P2, P3, ...) of the acute cerebral infarction patient can be received from the imaging device 100 via the communication unit 210 (S510). At this time, the imaging device 100 can acquire images (P1, P2, P3, ...) of the acute cerebral infarction patient. For example, the imaging device 100 can acquire magnetic resonance imaging (MRI) images of the acute cerebral infarction patient.

[0057] In the classification step, the processor 222 classifies images (P1, P2, P3, ...) of an acute cerebral infarction patient into a first image corresponding to a fluid-attenuated inversion recovery (FLAIR) image and a second image corresponding to a diffusion-weighted image (DWI) (S520).

[0058] In the acquisition step, the processor 222 acquires a first image corresponding to the classified water-suppressed image and a second image corresponding to the classified diffusion-weighted image (S530). In this case, the first image P4 corresponding to the water-suppressed image can gradually darken the area of ​​acute cerebral infarction as time passes after the onset of acute cerebral infarction. That is, the first image P4 corresponding to the water-suppressed image represents the time of the onset of acute cerebral infarction, and the degree of representation of the infarct area of ​​acute cerebral infarction varies with the elapsed time. Furthermore, the second image P5 corresponding to the diffusion-weighted image displays the area of ​​acute cerebral infarction within a short time after the onset of acute cerebral infarction, allowing the area of ​​acute cerebral infarction to be immediately identified. Therefore, the present disclosure can determine the time point of acute cerebral infarction based on the discrepancy between the first image P4 corresponding to the water-suppressed image and the second image P5 corresponding to the diffusion-weighted image. In this case, the image P6 may be a water-suppressed image of a healthy subject, and the image P7 may be a diffusion-weighted image of a healthy subject.

[0059] Here, the processor 222 can generate machine learning-based reference data based on the first image and the second image. That is, the processor 222 can generate machine learning model AIM-based reference data based on the first image data ID1 corresponding to a water-suppressed image and the second image data ID2 corresponding to a diffusion-weighted image. In this case, the first image and the second image may be of a patient with acute cerebral infarction whose onset time of acute cerebral infarction has not been confirmed.

[0060] In the matching step, the processor 222 can match the first and second images with Montreal Neurological Institute (MNI) regions corresponding to activated regions of the brain (S540). In the detection step, the processor 222 can detect infarct regions in the second image (S550). The processor 222 can also normalize the intensity of the second image, perform machine learning-based learning based on the standardized second image, and detect infarct regions in the second image learned and analyzed based on machine learning. The second image corresponding to such a standardized diffusion-weighted image can be output with improved image quality and a higher readability when learned and analyzed by the processor 222 based on the machine learning model AIM. The MNI regions can include MNI coordinate values ​​output as result values ​​for activated regions of the brain when analyzing the brain image. The infarct region may be a region where normal blood supply to growing tissue is impaired due to narrowing or blockage of blood vessels caused by thrombus, embolism, etc.

[0061] In the output step, the processor 222 performs learning based on the machine learning model AIM based on the first image data for the first image, the second image data for the second image, and the infarction region data for the infarction region, and outputs a probability prediction value for the time of occurrence of acute cerebral infarction that has been learned and analyzed based on the machine learning model AIM (S560).

[0062] In addition, the output step may involve the processor 222 learning the machine learning model AIM based on the first image data, the second image data, the infarction area data, and the clinical data of acute cerebral infarction, and further outputting a probability prediction value for the time point of occurrence of acute cerebral infarction learned and analyzed based on the machine learning model AIM (S560).

[0063] Furthermore, in the output step, the processor 222 can further label the area of ​​the infarct region, perform learning based on the machine learning model AIM based on the first image data, the second image data, and the area data of the infarct region, and further output a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning model AIM (S560).

[0064] In addition, the output step may further involve the processor 222 labeling the area of ​​the infarct region by location, performing learning based on the machine learning model AIM based on the first image data, the second image data, and the area data by location of the infarct region, and further outputting a probability prediction value for the time of occurrence of acute cerebral infarction that has been learned and analyzed based on the machine learning model AIM (S560).

[0065] Furthermore, in the output step, the processor 222 can further label the area of ​​the metastasis state by location of the infarct region, perform learning based on the machine learning model AIM based on the first image data, the second image data, and the area data of the metastasis state by location of the infarct region, and further output the probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning model AIM (S560).

[0066] At this time, the processor 222 combines the extracted image of the infarct region obtained by preprocessing with four-dimensional data and uses it as input data for a convolutional neural network (CNN) for learning. In the present disclosure, since a convolutional neural network is used, it is possible to learn information such as the location information and volume of the infarct region, which are lost when using an existing support vector machine or DNN.

[0067] In the output step, the processor 222 can output the result of the probability prediction value for the recommended occurrence time point of acute cerebral infarction corresponding to the input information of the input unit 230 (S560). The display unit 240 can display the probability prediction value for the occurrence time point of acute cerebral infarction output by the processor 222. For example, the display unit 240 can display the probability prediction values ​​belonging to each interval of the occurrence time point of acute cerebral infarction: 0 to 3 hours, 3 to 4.5 hours, 4.5 to 6 hours, 6 to 12 hours, and over 12 hours.

[0068] Meanwhile, a computer program stored in a computer-readable storage medium, when executed by one or more processors, can perform the following operations to perform the acute cerebral infarction occurrence time prediction method performed by the prediction device 200.

[0069] The operations may include receiving images of an acute cerebral infarction patient from the imaging device 100, classifying the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, and acquiring the first image and a second video.

[0070] Thereafter, the operations may include aligning the first image and the second image with MNI regions corresponding to activated brain regions, detecting the infarct region in the second image, and outputting a probability prediction value for the time of occurrence of acute cerebral infarction based on the first image, the second image, and the infarct region.

[0071] Here, the output operation can also perform machine learning-based learning based on the first image, the second image, and the infarct area, and output a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on machine learning.

[0072] In this case, the detection operation may standardize the intensity of the second image, perform machine learning-based learning based on the standardized second image, and detect the infarct region in the second image learned and analyzed based on machine learning. Here, the output operation may label the area of ​​the infarct region, perform machine learning-based learning based on the first image, the second image, and the area of ​​the infarct region, and output a probability prediction value for the time point of occurrence of acute cerebral infarction learned and analyzed based on machine learning.

[0073] The operations may further include generating machine learning-based reference data based on the first image and the second image.

[0074] Meanwhile, the present disclosure can determine whether the golden hour (4.5 hours) during which a thrombolytic agent, one of the treatment methods for cerebral infarction, can be administered has elapsed. In this case, the elapsed time of patients in existing images for learning can be classified into a first group for those less than 4.5 hours and a second group for those 4.5 hours or more. In this case, the present disclosure can utilize not only the images for learning but also the classified elapsed time of patients as learning data.

[0075] When generating machine learning-based reference data based on first and second images of multiple patients, the present disclosure can classify the multiple patients into a training group, a validation group, and a test group, and perform machine learning-based learning using the first and second images of patients included in the training group. Furthermore, the present disclosure can determine the learning form of the machine learning-based system through the validation group, and can determine whether the test group is applied to a machine learning model and set as reference data.

[0076] The present disclosure may extract first feature information from a first image. Here, the first feature information may be one or more feature elements extracted from a specific region of the image data. In this case, the present disclosure may extract the first feature information by applying at least one of a Gray Level Cooccurrence Matrix (GLCM), a Run-Length Matrix (particularly, a Gray Level Run-Length Matrix (GLRLM)), and a Local Binary Pattern (LBP) to the signal intensity and gradient of the infarct region.

[0077] The present disclosure can extract second feature information that can be acquired from the infarct region in the second image, where the second feature information can include size or volume information of the infarct region, and can include mean, standard deviation, skewness, and kurtosis calculated for each of signal intensity, signal gradient, and LBP (Local Binary Pattern) map in the infarct region.

[0078] In this way, the present disclosure can provide information that allows medical staff to safely and quickly determine the time point at which acute cerebral infarction occurs.

[0079] 1 and 2, at least one component may be added or removed depending on the performance of the components. Furthermore, it is easily understood by those skilled in the art that the relative positions of the components may be changed depending on the performance or structure of the system.

[0080] Although FIG. 5 shows multiple steps being performed sequentially, this is merely an illustrative example of the technical concept of this embodiment, and a person having ordinary skill in the art to which this embodiment pertains can change the order shown in FIG. 5 or perform one or more of the multiple steps in parallel within the scope of the essential characteristics of this embodiment, and therefore, FIG. 5 is not limited to a chronological order.

[0081] Meanwhile, the disclosed embodiments may be realized in the form of a recording medium storing computer-executable instructions. The instructions may be stored in the form of program code, which, when executed by a processor, generates program modules to perform the operations of the disclosed embodiments. The recording medium may be realized as a computer-readable recording medium.

[0082] Computer-readable recording media include all types of recording media that store computer-readable instructions, such as ROM (Read Only Memory), RAM (Random Access Memory), magnetic tape, magnetic disk, flash memory, and optical data storage devices.

[0083] The disclosed embodiments have been described above with reference to the accompanying drawings. Those skilled in the art will understand that the present disclosure may be embodied in forms different from the disclosed embodiments without changing the technical concept or essential features of the present disclosure. The disclosed embodiments are illustrative and should not be construed as limiting.

Claims

1. a communication unit that communicates with an imaging device that captures images of patients with acute cerebral infarction; a processor for controlling operations related to predicting the time of occurrence of acute cerebral infarction; Including, The processor: receiving images of the patient with acute cerebral infarction from the imaging device via the communication unit, classifying the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, and acquiring the first image and a second moving image; aligning the first and second images with MNI regions corresponding to activated regions of the brain to detect infarct regions in the second image; An apparatus for predicting the time of occurrence of acute cerebral infarction, comprising: an apparatus for predicting the time of occurrence of acute cerebral infarction, the apparatus outputting a probability prediction value for the time of occurrence of acute cerebral infarction based on the first image, the second image, and the infarct region.

2. The processor: performing machine learning-based learning based on the first image, the second image, and the infarct region; The apparatus for predicting the time of occurrence of acute cerebral infarction according to claim 1, characterized in that it outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

3. The processor: normalizing the intensity of the second image and performing machine learning-based training based on the normalized second image; The apparatus for predicting the occurrence time of acute cerebral infarction according to claim 2, characterized in that an infarct area is detected in the second image that has been learned and analyzed based on the machine learning.

4. The processor: Labeling the area of ​​the infarct region; performing machine learning-based learning based on the first image, the second image, and the area of ​​the infarct region; The apparatus for predicting the time of occurrence of acute cerebral infarction according to claim 3, characterized in that it outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

5. The processor: The apparatus for predicting the onset time of acute cerebral infarction according to claim 2, further generating reference data for the machine learning platform based on the first image and the second image.

6. In a method for predicting the occurrence time of acute cerebral infarction performed by a prediction device, receiving images of the patient with acute cerebral infarction from an imaging device by a communication unit of the prediction device; classifying the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image by a processor of the prediction device; acquiring, by the processor, the first image and the second image; aligning, by the processor, the first and second images with MNI regions corresponding to activated regions of the brain; detecting, by the processor, an infarct region in the second image; outputting, by the processor, a probability prediction value for the time point of occurrence of the acute cerebral infarction based on the first image, the second image, and the infarct region; A method comprising:

7. The step of outputting the probability prediction value for the time point of occurrence of acute cerebral infarction includes: The method according to claim 6, characterized in that the processor performs machine learning-based learning based on the first image, the second image, and the infarct area, and outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

8. Detecting an infarct region in the second image includes: The method of claim 7, wherein the processor normalizes the intensity of the second image, performs machine learning-based learning based on the normalized second image, and detects the infarct area in the second image that has been learned and analyzed based on the machine learning.

9. The step of outputting the probability prediction value for the time point of occurrence of acute cerebral infarction includes: The method according to claim 8, characterized in that the processor labels the area of ​​the infarct region, performs machine learning-based learning based on the first image, the second image, and the area of ​​the infarct region, and outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

10. The method of claim 7 , further comprising generating, by the processor, reference data for the machine learning base based on the first image and the second image.

11. an imaging device for acquiring images of patients with acute cerebral infarction; a prediction device for the occurrence of acute cerebral infarction that communicates with the imaging device; Including, The prediction device includes: receiving images of the patient with acute cerebral infarction from the imaging device, classifying the images into a first image corresponding to a water-suppressed image and a second image corresponding to a diffusion-weighted image, and acquiring the first image and a second moving image; aligning the first and second images with MNI regions corresponding to activated regions of the brain to detect infarct regions in the second image; A system for predicting the time of occurrence of acute cerebral infarction, characterized in that it outputs a probability prediction value for the time of occurrence of the acute cerebral infarction based on the first image, the second image, and the infarction area.

12. The prediction device includes: performing machine learning-based learning based on the first image, the second image, and the infarct region; The system for predicting the time of occurrence of acute cerebral infarction according to claim 11, characterized in that it outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

13. The prediction device includes: normalizing the intensity of the second image and performing machine learning-based training based on the normalized second image; The system for predicting the occurrence time of acute cerebral infarction according to claim 12, characterized in that an infarct area is detected in the second image that has been learned and analyzed based on the machine learning.

14. The prediction device includes: Labeling the area of ​​the infarct region; performing machine learning-based learning based on the first image, the second image, and the area of ​​the infarct region; The system for predicting the time of occurrence of acute cerebral infarction according to claim 13, characterized in that it outputs a probability prediction value for the time of occurrence of acute cerebral infarction learned and analyzed based on the machine learning.

15. The prediction device includes: The system for predicting the occurrence time of acute cerebral infarction according to claim 12, further comprising generating reference data for the machine learning platform based on the first image and the second image.

Images