Method and apparatus for detecting microembolic signals among doppler ultrasound signals using artificial intelligence

Abstract

An aspect of the present invention provides a method for providing information about whether a microembolism is present in a subject by using Doppler ultrasound. The method comprises the steps of: acquiring a first Doppler ultrasound measurement result by using a first probe on a measurement target part of the subject; pre-processing the first Doppler ultrasound measurement result by means of a processor to acquire a pre-processed first Doppler ultrasound measurement result; determining whether an MES is present from the pre-processed first Doppler ultrasound measurement result by using a first artificial intelligence algorithm by means of the processor; and providing information about the microembolism in the subject by means of the processor according to whether the MES is present.

Description

Method and device for detecting microembolism signals from Doppler ultrasound signals using artificial intelligence

[0001] The present invention relates to a method and device for detecting microembolic signals (MES) from Doppler ultrasound signals using artificial intelligence.

[0002]

[0003] An embolus is a substance that travels through the bloodstream and blocks a blood vessel. It can take various forms, including blood clots, fatty tissue, air bubbles, or other foreign substances. If an embolus blocks a blood vessel, it can lead to serious conditions such as stroke, pulmonary embolism, or myocardial infarction.

[0004] Microembolism (MEs) are extremely small and initially asymptomatic. However, their continued accumulation or migration can potentially lead to fatal long-term complications. These microembolisms appear as MESs on Doppler ultrasound images and are used as early predictors of embolism, vascular occlusion, and cancer metastasis.

[0005] Existing Doppler ultrasound devices are bulky, heavy, expensive, and require skilled medical personnel, making them difficult for patients to use individually. To overcome these limitations, the development of highly portable Doppler ultrasound devices that can be easily used by non-specialists is needed.

[0006]

[0007] The present invention aims to provide a method and system for detecting MES from a Doppler ultrasound signal using artificial intelligence.

[0008]

[0009] A representative configuration of the present invention to achieve the above purpose is as follows.

[0010] According to one aspect of the present invention, a method for providing information on whether a microembolism exists in a subject using Doppler ultrasound is provided, the method comprising: obtaining a first Doppler ultrasound measurement result by using a first probe on a measurement target portion of the subject; preprocessing the first Doppler ultrasound measurement result by a processor to obtain a preprocessed first Doppler ultrasound measurement result; determining, by the processor, whether a microembolic signal (MES) exists from the preprocessed first Doppler ultrasound measurement result by using a first artificial intelligence algorithm; and providing, by the processor, information on a microembolic signal in the subject based on whether the MES exists.

[0011] According to another aspect of the present invention, an MES detection device is provided, including a sensor module including a first probe, a storage storing a first artificial intelligence algorithm, and a processor for obtaining a first Doppler ultrasound measurement result by using the first probe on a measurement target portion of the object, preprocessing the first Doppler ultrasound measurement result to obtain a preprocessed first Doppler ultrasound measurement result, determining whether an MES exists from the preprocessed first Doppler ultrasound measurement result using the first artificial intelligence algorithm, and providing information on a microembolism in the object based on the presence or absence of the MES.

[0012] In addition, a non-transitory computer-readable recording medium recording another method for implementing the present invention, another device, and a computer program for executing the method are further provided.

[0013]

[0014] According to the present invention, the presence or absence of MES can be quickly and accurately determined using artificial intelligence from a signal acquired through Doppler ultrasound, thereby contributing to the early diagnosis and prevention of diseases related to microembolism.

[0015] According to the present invention, by combining a portable and easy-to-use Doppler ultrasound device with artificial intelligence technology, even ordinary people, not medical professionals, can easily utilize it to perform continuous health monitoring at home.

[0016]

[0017] FIG. 1 is a block diagram of an MES detection device according to one embodiment of the present invention.

[0018] FIG. 2 is a flowchart showing an operation method of an MES detection device according to one embodiment of the present invention.

[0019] FIG. 3 is a flowchart showing a preprocessing execution step in an operating method of an MES detection device according to one embodiment of the present invention.

[0020] Figure 4 illustrates a learning method of an artificial intelligence algorithm used in an operating method of an MES detection device according to one embodiment of the present invention.

[0021] FIG. 5 is a flowchart illustrating a method of operating an MES detection device according to one embodiment of the present invention for changing an amplitude threshold value and a duration threshold value as MES is detected.

[0022] FIG. 6 is a flowchart showing an operation method of an MES detection device having a blood vessel location recognition function according to one embodiment of the present invention.

[0023]

[0024] The following detailed description of the present invention refers to the accompanying drawings, which illustrate specific embodiments in which the present invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the present invention. It should be understood that the various embodiments of the present invention, while different from each other, are not necessarily mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be modified and implemented from one embodiment to another without departing from the spirit and scope of the present invention. Furthermore, it should be understood that the positions or arrangements of individual components within each embodiment may also be modified without departing from the spirit and scope of the present invention. Accordingly, the following detailed description is not to be taken in a limiting sense, and the scope of the present invention is to be construed to encompass the scope of the claims and all equivalents thereof. Like reference numerals in the drawings represent the same or similar elements throughout the several aspects.

[0025] While terms such as "first" and "second" may be used to describe various components, these components should not be limited by these terms. These terms are used solely to distinguish one component from another. For example, without departing from the scope of the present invention, a first component could be referred to as a "second component," and similarly, a second component could also be referred to as a "first component." The term "and / or" encompasses any combination of multiple related items described herein or any one of multiple related items described herein.

[0026] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0027] The terminology used in this application is only used to describe specific embodiments and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, it should be understood that the terms "comprise" or "have" indicate the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but do not exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0028] In this regard, the terms "about," "substantially," etc., used throughout the specification, are used in a sense of numerical value or a value close to that numerical value when manufacturing and material tolerances inherent to the meanings mentioned are presented, and are used to prevent unscrupulous infringers from unfairly exploiting the disclosure, which contains precise or absolute numerical values ​​to aid understanding of the present invention. The terms "step of doing ~" or "step of ~" used throughout the specification of the present invention do not mean "step for ~."

[0029] In this specification, the term "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Furthermore, a single unit may be realized using two or more pieces of hardware, and two or more units may be realized by a single piece of hardware.

[0030] Some of the operations or functions described herein as being performed by a terminal, apparatus, or device may instead be performed by a server connected to the terminal, apparatus, or device. Similarly, some of the operations or functions described herein as being performed by a server may also be performed by a terminal, apparatus, or device connected to the server.

[0031] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. Terms defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and will not be interpreted in an idealized or overly formal sense unless explicitly defined herein.

[0032] Hereinafter, with reference to the attached drawings, preferred embodiments of the present invention will be described in more detail. In order to facilitate an overall understanding in describing the present invention, identical reference numerals will be used for identical components in the drawings, and redundant descriptions of identical components will be omitted.

[0033] FIG. 1 is a block diagram of an MES detection device according to one embodiment of the present invention.

[0034] Referring to FIG. 1, the MES detection device (100) may include a sensor module (110), a processor (120), a motion control module (130), and storage (140). Depending on the embodiment, the MES detection device (100) may be implemented as a wearable device or a surgical device. Depending on the embodiment, the MES detection device (100) may be performed in real time.

[0035] In some embodiments, the sensor module (110) may include a probe. One or more probes may be used, and each probe may be utilized for a specific purpose. For example, the sensor module (110) may include a first probe (112) and / or a third probe (114).

[0036] The first probe (112) may be focused on analyzing ultrasonic reflection signals generated from the subject's blood vessels to determine the presence of microembolism. This allows for real-time detection of microembolism within the bloodstream, contributing to the early detection of diseases such as embolism and stroke.

[0037] The third probe (114) can be used to identify the location of the target blood vessel. Blood vessel structures can vary depending on the anatomical differences of each patient, and precise identification of the blood vessel location is essential for accurate MES detection. The third probe (114) can be used to determine the relative locations of major blood vessels based on blood flow velocity and direction, and based on this, can assist the first probe (112) in collecting signals at the optimal measurement point. By utilizing these multiple probes, the present invention enables more precise blood flow analysis and minimizes measurement errors that can occur with existing single probes. This provides medical staff with more accurate diagnostic information, and enables more effective patient health management through real-time monitoring.

[0038] In one embodiment, the Doppler ultrasound analysis results may be provided as spectral Doppler ultrasonography. Spectral Doppler ultrasonography is a method for evaluating blood flow velocity by graphically measuring it and then quantifying it. The analysis results of spectral Doppler ultrasound may also be provided in the form of a Doppler signal waveform. The probe can output information on blood flow velocity and direction in the form of a waveform over time, enabling quantitative analysis of the characteristics of blood flow within a blood vessel. In an embodiment of the present invention, the frequency spectrum of the Doppler waveform is analyzed to extract a pattern of blood flow velocity variation within a blood vessel, and the location of a blood vessel can be determined by searching for a location where a specific flow pattern remains constant. By analyzing the waveform data, characteristics such as blood flow velocity, pulsatility, and turbulence within a blood vessel can be precisely evaluated, thereby enabling more accurate detection of the location of the blood vessel.

[0039] In one embodiment, the Doppler ultrasound analysis results may be provided as color Doppler ultrasonography. Color Doppler ultrasonography is a Doppler examination method that assigns colors to indicate the direction of blood flow and determines the brightness of the colors according to the velocity of the blood flow. Color Doppler ultrasonography can also provide an image of the space to be measured. The probe can visually represent the relative positions of blood vessels within a specific space based on blood flow, and can perform a function of searching for blood vessel structures in 2D or 3D space based on the ultrasound probe. In an embodiment of the present invention, the searched blood vessel locations can be expressed in the form of color differences (color mapping) or velocity vectors (vector mapping), and specific blood vessels can be effectively identified by emphasizing areas with high blood flow velocities. Furthermore, these image-based analysis results can be utilized to automatically recognize blood vessel structures and select blood vessels to be measured using machine learning and artificial intelligence models.

[0040] In an embodiment of the present invention, MES can be detected and the location of a blood vessel to be measured can be determined by combining the results of a tone Doppler ultrasound examination and the results of a color Doppler ultrasound examination.

[0041] In this specification, the names "first probe (112)" and "third probe (114)" are merely formulas used to distinguish their respective functions, and they do not necessarily have to be different devices. The same probe may perform both MES detection and target vessel location identification. Furthermore, multiple probes performing the same purpose may be mounted on a single MES detection device. For example, multiple probes may be positioned for the same measurement purpose to increase the reliability of MES detection, and multiple probes may be utilized to enable measurements from various angles to increase the precision of vessel location identification.

[0042] Depending on the embodiment, the probe may be implemented to be embedded in or coupled to the MES detection device (100).

[0043] The probe transmits ultrasound signals and receives reflected signals, analyzing the signals reflected from red blood cells in the bloodstream to measure blood flow velocity. The probe measures blood flow velocity using the Doppler effect and has a built-in ultrasound transducer. Typically, frequencies in the 1-10 MHz range are used. In one embodiment, an ultrasound frequency of 2 MHz can be used to measure cerebral blood flow.

[0044] According to an embodiment, one or more probes may be mounted on the MES detection device (100), and measurement may be performed by selecting an optimal probe based on the relative position with respect to the blood vessel to be measured. In one embodiment, two probes may be mounted on the left and right sides of the MES detection device (100), and MES measurement may be performed using the probe close to the blood vessel to be measured. In one embodiment, two probes may be mounted on the left and right sides of the MES detection device (100), and MES measurement may be performed using the probe close to the blood vessel to be measured. It may be configured in a manner that one probe performs blood vessel location search, and the other probe performs MES detection. Signals may be collected from various angles by utilizing multiple probes, and reliability may be increased. The quality of data collected during measurement may be analyzed, and if the signal strength of a specific probe is low or distorted, it may be switched to another probe.

[0045] According to an embodiment, the MES detection device (100) may include a support or a fixing band. The support does not completely encircle the measurement target area, but is structured to be fixed or attached only to a portion of the measurement target area in the form of a headband. The support may be relatively lightweight and designed to be stably mounted at a specific area. The fixing band may have a structure that completely encircles the measurement target area in the form of a hairband and may be designed to more firmly fix the probe. The fixing band may include an elastic material. Both the support and the fixing band may include an adjustable fastening device to allow the probe to change its position and angle during measurement.

[0046] Depending on the embodiment, the measurement target site may be the head, neck, arm, wrist, chest, thigh, or ankle. The measurement target site can be selected based on the location of major blood vessels with active blood flow, and each site can be selected based on the measurement purpose and application environment. For the head site, transcranial Doppler (TCD) ultrasound can be used to measure blood flow within the cerebral blood vessels and detect microembolic signals (MES) in the middle cerebral artery (MCA) and internal carotid artery. For the neck site, carotid artery ultrasound can be used to analyze blood flow status and stenosis within the carotid artery, and can be measured so that MES can be detected in the carotid artery before traveling to the brain. For the arm and wrist sites, blood flow in peripheral arteries can be measured, which can be used to assess the health of systemic blood circulation. In particular, blood flow analysis is possible in the radial artery and brachial artery. For the chest area, it can be used to monitor blood flow in the major arteries that connect the chest to the heart (e.g., the aorta, subclavian artery, etc.). For the thigh area, blood flow can be measured in the femoral artery to analyze the blood flow status of the lower extremities, which can be useful for detecting arterial occlusion and vascular disease in the lower extremities. For the ankle area, blood flow can be measured to detect peripheral artery disease (PAD), and the ankle-brachial index (ABI) can be calculated to assess blood flow problems in the lower extremities. Furthermore, the measurement method at each site can vary depending on the probe placement and fixation method. For example, a support or fixation band can be used to stabilize the probe position at the head and neck areas, while an elastic fixation band can be used at the wrist and ankle areas to minimize movement during measurement.

[0047] The position and angle of the probe can be adjusted by the motion control module (130), enabling optimal ultrasound measurements of the target blood vessel. Additionally, when the target blood vessel changes or the patient moves, the direction of the probe is automatically adjusted for optimal ultrasound measurements.

[0048] According to an embodiment, the target blood vessel to be measured may include the middle cerebral artery (MCA), the internal carotid artery (ICA), the external carotid artery (ECA), the common carotid artery (CCA), the brachial artery, the subclavian artery, the internal thoracic artery, the axillary artery, the femoral artery, the deep femoral artery, the radial artery, the ulnar artery, the subclavian artery, the aorta, the tibial artery, or the dorsalis pedis artery.

[0049] The target vessel location can be a relative position from the probe. This relative position can be set based on the probe's coordinate system and can be expressed in one of three commonly used coordinate systems: the Cartesian coordinate system, the polar coordinate system, and the cylindrical coordinate system. When using multiple probes, the relative position can be determined based on the position of one of these probes or their center points.

[0050] In another embodiment, the position and angle of the probe within the MES detection device (100) may be fixed. The position and angle of the probe may be fixed to an optimal state for MES measurement.

[0051] According to an embodiment, the processor (120) may include a CPU (Central Processing Unit) (122) and an NPU (Neural Processing Unit) (124). The CPU (122) may control the overall operation of the MES detection device (100). The CPU (122) may include one processor core (Single Core) or multiple processor cores (Multi-Core). The CPU (110) may process or execute programs and / or data stored in the storage (300). For example, the CPU (110) may control the function of the NPU (124) by executing programs and / or modules stored in the storage (300). CPU (122) can also implement a neural network model using languages ​​such as Java, C / C++, Python, R, and implementation languages ​​such as Tensorflow, Keras, and Pytorch based on Python.

[0052] The processor (120) can store the Doppler ultrasound measurement results obtained through the sensor module (110) in storage. The processor (120) can adjust the position and angle of the probe to obtain an optimal signal by considering the position of the blood vessel to be measured.

[0053] The processor (120) can be controlled to perform preprocessing on the Doppler ultrasound measurement results obtained through the sensor module (110). The processor (120) can prevent data loss by converting the obtained Doppler ultrasound measurement results into continuous frames. The processor (120) applies a filtering technique to remove low-frequency or high-frequency noise and improve signal quality. The processor (120) can convert the signal to emphasize specific patterns (amplitude, duration, etc.) to make the characteristics of the microembolism signal (MES) more distinct.

[0054] The processor (120) can analyze the location of blood vessels based on Doppler ultrasound measurement results to determine relative coordinates. The processor (120) can automatically search for target blood vessels by considering the flow velocity and direction of the blood vessels. The processor (120) can dynamically adjust the position and angle of the probe based on the blood vessel search results through the motion control module (130) to maintain an optimal measurement environment.

[0055] The processor (120) can extract the feature vector (amplitude, duration, frequency fluctuation, etc.) of the signal analysis signal for MES detection to evaluate the possibility of a microembolism signal. The processor (120) can classify the signal by applying a signal analysis model based on the first artificial intelligence algorithm (142) and determine the presence or absence of MES. Based on the analysis results, the processor (120) can compare signal intensities and patterns to evaluate the reliability of MES detection.

[0056] The processor (120) can update the amplitude threshold and / or duration threshold used in the preprocessing process based on the MES detection results. The processor (120) can dynamically adjust the amplitude threshold and duration threshold based on the detected MES pattern.

[0057] The processor (120) can perform dynamic threshold value change in the preprocessing process. When the presence of MES is determined from the Doppler ultrasound measurement results using the first artificial intelligence algorithm (142), the processor (120) can lower the first amplitude threshold value (146), raise the first duration threshold value (148), or lower the first amplitude threshold value (146) and raise the first duration threshold value (148). The processor (120) can apply the changed first amplitude threshold value (146), the changed second duration threshold value (148), or the changed first amplitude threshold value (146) and the changed second duration threshold value (148) in the step of preprocessing the Doppler ultrasound measurement results to be performed thereafter and obtaining a preprocessed Doppler ultrasound measurement result. The processor (120) can adjust the detection sensitivity by applying the updated threshold value to the new measurement result.

[0058] The processor (120) may provide information about microembolism within a subject based on the presence or absence of MES. In one embodiment, the information about microembolism may include the presence or absence of microembolism, the location where microembolism is detected, the detection frequency of microembolism, the signal intensity of MES, or a combination thereof. In one embodiment, the processor (120) may store the information about microembolism in real time and output the analysis results. The processor (120) may support analyzing changes over time by recording a data log. In one embodiment, the processor (120) may provide information about microembolism within a subject in real time. In one embodiment, the information about microembolism may be used to predict cerebral infarction, myocardial infarction, or cancer metastasis.

[0059] In one embodiment, the processor (120) can perform machine learning of an artificial intelligence algorithm through the NPU (124). The NPU (124) is a processor designed to optimize machine learning operations, including deep learning, and can accelerate the learning and execution of artificial intelligence algorithms. In particular, it provides high computational efficiency in neural network models requiring a large amount of matrix operations, and can support faster computation speeds than the CPU (122) through parallel processing. In one embodiment, the NPU (124) can be used to perform learning and optimized inference of the first artificial intelligence algorithm (142) and the third artificial intelligence algorithm (144) for analyzing Doppler ultrasound signals.

[0060] The processor (120) can be utilized not only to detect MES from Doppler ultrasound signals, but also to identify the location of a target blood vessel. The processor (120) analyzes the probe's signal during the process of searching for the location of a target blood vessel to be measured to evaluate blood flow velocity and direction, and can more precisely determine the relative location of a blood vessel by rapidly processing a large amount of signal data. Through this, the processor (120) can automatically select a probe close to the target blood vessel to be measured, and maintain optimal blood flow search conditions to improve the accuracy of MES detection.

[0061] The processor (120) receives filtered Doppler ultrasound data as input during the signal preprocessing process, and performs artificial intelligence-based feature extraction and signal pattern classification to more precisely analyze the characteristics of the signal. Furthermore, the processor (120) may be used in the learning process of a machine learning model for blood vessel location search and MES detection, and during real-time data analysis, the presence or absence of MES can be quickly determined by performing inference based on the learned model.

[0062] In one embodiment, the CPU (122) and NPU (124) may be appropriately combined and utilized depending on the need for real-time analysis and computational performance requirements. That is, the CPU alone can be used to analyze signals and detect MES by executing a learned model. However, when real-time processing or large-scale parallel computation is required, the NPU (124) may be additionally utilized. Through this, the present invention can optimize computational performance while enhancing the accuracy of neural network-based MES detection, thereby enabling rapid and efficient diagnosis.

[0063] According to an embodiment, the NPU (124) may create a neural network, train or learn a neural network, perform a calculation based on training data, generate an information signal based on the result of the calculation, or retrain the neural network. The NPU (124) is not limited to the name, and may be a processor for executing a neural network, such as a GPU or TPU.

[0064] According to various embodiments, the neural network models may include various types of models such as CNN (Convolution Neural Network), R-CNN (Region with Convolution Neural Network), RPN (Region Proposal Network), RNN (Recurrent Neural Network), S-DNN (Stacking-based deep Neural Network), S-SDNN (State-Space Dynamic Neural Network), Deconvolution Network, DBN (Deep Belief Network), RBM (Restricted Boltzman Machine), Fully Convolutional Network, LSTM (Long Short-Term Memory) Network, Classification Network, etc., but are not limited to the above-mentioned models.

[0065] Depending on the embodiment, the neural network models of the reinforcement learning agent may include various models such as DQN (Deep Q-Network), DDQN (Double DQN), Dueling DQN, DDPG (Deep Deterministic Policy Gradient), and DDDQN (Dueling Double DQN), but are not limited to the above-described models.

[0066] In some embodiments, the processor (120) may utilize separate storage (140) for storing programs corresponding to neural network models. The processor (120) may further include separate intellectual property (IP) blocks for processing the numerous operations required to operate the neural network. For example, the separate IP blocks may further include a graphical processing unit (GPU) or an accelerator for quickly performing specific operations.

[0067] According to an embodiment, the motion control module (130) may control the movement of the MES detection device (100). For example, the motion control module (130) may control the configuration (e.g., the shape of the support, the length of the support, the tightening of the fixing band, the position of the probe on the support and / or the fixing band, the angle of the probe) for moving the MES detection device (100) according to the location of the blood vessel to be measured.

[0068] In one embodiment, the motion control module (130) may adjust the position and angle of the first probe (112) or the third probe (114) to direct the first probe (112) or the third probe (114) toward the location of the blood vessel to be measured. In one embodiment, the processor (120) may adjust the position and angle of the first probe (112) or the third probe (114) through the motion control module (130) based on information about the location of the blood vessel to be measured. In one embodiment, the motion control module (130) may adjust the shape of the support, the length of the support, or the tightness of the fixing band.

[0069] According to an embodiment, the storage (140) may be a storage location for storing data, and may store an operating system (OS), various programs, and various data. For example, the storage (140) may correspond to a non-volatile memory. The storage (140) may include a read-only memory (ROM), a flash memory, a phase-change RAM (PRAM), a magnetic RAM (MRAM), a resistive RAM (RRAM), a ferroelectric RAM (FRAM), and the like. According to an embodiment, the storage (140) may be implemented as a hard disk drive (HDD), a solid state drive (SSD), and the like.

[0070] According to an embodiment, the storage (140) may store a first artificial intelligence algorithm (142) and a third artificial intelligence algorithm (144). The first artificial intelligence algorithm (142) is a neural network model for determining the presence or absence of a microembolic signal (MES) in a preprocessed Doppler ultrasound measurement result. The first artificial intelligence algorithm (142) may receive the preprocessed Doppler ultrasound measurement result as input, analyze the pattern of the MES, and provide a detection result. The third artificial intelligence algorithm (144) is an algorithm for searching the location of a target blood vessel and determining relative coordinates, and may be used to automatically search the target blood vessel and set an optimal probe position. These artificial intelligence algorithms may be stored as pre-learned models or designed to be continuously updated when new data is added.

[0071] According to an embodiment, the storage (140) may store a first amplitude threshold value (146) and a first duration threshold value (148). The first amplitude threshold value (146) is a reference value for removing noise based on the amplitude of the signal, and may be used to identify only signals having an amplitude higher than a certain level in a Doppler ultrasound signal as MES candidates. The first duration threshold value (148) is a reference value for determining only signals that last longer than a specific period of time as MES, and filters out instantaneous changes in the signal and enables reliable MES detection. These threshold values ​​may be dynamically updated based on data stored in the storage (140) and real-time analysis results, and may also be adjusted according to user settings to improve detection accuracy.

[0072] In addition, the storage (140) may store an initial learning dataset, user-customized threshold settings, patient-specific measurement history, raw data of Doppler ultrasound signals, and analyzed feature vectors. This allows for verification of the reliability of MES detection results by comparing them with previously measured data, and supports long-term monitoring of changes in blood flow patterns of specific patients. Furthermore, the storage (140) may be linked to a cloud-based repository, enabling real-time diagnostic data sharing through connection to a remote medical system, or allowing medical staff to retrieve past data and perform in-depth analysis when necessary.

[0073] In one embodiment, the MES detection device may include a sensor module including a first probe, a storage storing a first artificial intelligence algorithm, and a processor that obtains a first Doppler ultrasound measurement result by using the first probe on a measurement target portion of the object, preprocesses the first Doppler ultrasound measurement result to obtain a preprocessed first Doppler ultrasound measurement result, determines whether an MES exists from the preprocessed first Doppler ultrasound measurement result using the first artificial intelligence algorithm, and provides information on a microembolism in the object based on the presence or absence of the MES.

[0074] In one embodiment, the MES detection device may include a sensor module including a first probe, a storage storing a first artificial intelligence algorithm and a third artificial intelligence algorithm, and a processor that obtains a third Doppler ultrasound measurement result by using the third probe on a measurement target portion of a subject, preprocesses the third Doppler ultrasound measurement result to obtain a preprocessed third Doppler ultrasound measurement result, obtains a measurement target blood vessel location from the preprocessed third Doppler ultrasound measurement result using the third artificial intelligence algorithm, orients the first probe toward the measurement target blood vessel location of the measurement target portion of the subject, preprocesses the first Doppler ultrasound measurement result to obtain a preprocessed first Doppler ultrasound measurement result, determines whether an MES is present from the preprocessed first Doppler ultrasound measurement result using the first artificial intelligence algorithm, and provides information on a microembolism in the subject based on the presence or absence of the MES.

[0075] FIG. 2 is a flowchart showing an operation method of an MES detection device according to one embodiment of the present invention.

[0076] The operating method of the MES detection device of the present invention includes the steps of obtaining a first Doppler ultrasound measurement result through a sensor module (S110), performing preprocessing of the obtained data (S120), determining the presence of MES using a first artificial intelligence algorithm (S130), and providing information on microembolism within the target object (S140). This operating method can be performed in real time.

[0077] In the step of obtaining the first Doppler ultrasound measurement result through the sensor module (S110), the first probe is used on the measurement target area and is used for MES detection.

[0078] In the step of performing preprocessing of acquired data (S120), the data is preprocessed to enable effective use of the first artificial intelligence algorithm. In one embodiment, the processor (120) performs noise removal and signal normalization.

[0079] In the step of determining the presence or absence of an MES using the first artificial intelligence algorithm (S130), the first artificial intelligence algorithm can detect an MES from a preprocessed Doppler ultrasound signal. In one embodiment, the first artificial intelligence algorithm can be trained based on MES clinical data. The MES clinical data may be clinical data labeled with a Doppler ultrasound measurement result and the presence or absence of an MES. The first artificial intelligence algorithm may be trained to match a Doppler ultrasound measurement result (e.g., a second Doppler ultrasound measurement result) preprocessed based on the MES clinical data with the presence or absence of an MES labeled based on the MES clinical data. In one embodiment, in the process of training the first artificial intelligence algorithm, the second Doppler ultrasound measurement result, which is obtained by preprocessing the Doppler ultrasound measurement result based on the MES clinical data, may be used as training data, and the clinical data on the presence or absence of an MES may be used as labeling data. In one embodiment, features (e.g., signal amplitude, frequency variation, duration, etc.) for the MES may be directly extracted and machine learning may be performed based on the extracted features.

[0080] In the step of providing information on microembolism within a subject (S140), information on microembolism within the subject may be provided based on the presence or absence of MES. According to one embodiment, the information on microembolism within the subject may include the presence or absence of microembolism, the detection frequency of microembolism, the signal intensity of MES, or a combination thereof. Dynamic thresholds may be included as needed. For example, when MES is detected, the first amplitude threshold (146) and the first duration threshold (148) may be dynamically changed, and the updated thresholds may be applied in a subsequent detection process. In addition, the detected data may be stored in the storage (140) and utilized as data for long-term patient monitoring and further analysis.

[0081] FIG. 3 is a flowchart showing a preprocessing execution step in an operating method of an MES detection device according to one embodiment of the present invention.

[0082] The step of performing preprocessing of acquired data (S120) in the operating method of the MES detection device of the present invention may include the step of continuously acquiring the first Doppler ultrasound measurement results as a plurality of Doppler ultrasound images (S210), the step of extracting a single or multiple signals from the plurality of Doppler ultrasound images (S220), the step of filtering a signal having an amplitude lower than a first amplitude threshold value in the single or multiple signals (S230), and / or the step of additionally filtering a signal having a duration longer than a first duration threshold value in the filtered single or multiple signals (S240).

[0083] In the step of continuously acquiring the first Doppler ultrasound measurement results as multiple Doppler ultrasound images (S210), a process of acquiring a blood flow signal of a target area of ​​measurement using a probe may be performed. In an embodiment of the present invention, the first probe is used for MES detection and measures continuous blood flow signals. During the signal acquisition process, multiple Doppler ultrasound images may be continuously collected, and data may be acquired in an overlapping manner so that the signal can be maintained continuously without data loss. The Doppler ultrasound signal is captured as continuous image frames, and may be segmented into a state where a certain portion overlaps to prevent signal loss. In one embodiment, the Doppler ultrasound signal is segmented into time intervals of 100 to 200 ms, and an overlap of 50% or more may be applied between each segment to prevent continuous data loss. In one embodiment, the segmented Doppler ultrasound signal may be overlapped by 10% to 90% or more between each segment to prevent continuous data loss. This enables continuous detection without signal loss even if the MES exists beyond the boundaries of a single frame.

[0084] In the step of extracting a single or multiple signals from multiple Doppler ultrasound images (S220), continuously acquired signals are arranged in time order, and frames are merged based on overlapping data with previous segments to extract a single or multiple signals that maintain a flow. The signal amplitude and duration can be adjusted to enable consistent analysis in all measurement environments by standardizing them. In one embodiment, the signal can be converted to the frequency domain using a Fourier transform, and the frequency spectrum can be analyzed to remove unnecessary noise. In particular, since the MES signal exhibits a pattern in a specific frequency band, the MES candidate signal can be more clearly distinguished through signal filtering and frequency analysis.

[0085] In the step of filtering signals having an amplitude lower than a first amplitude threshold value in a single or multiple signals (S230), a high-pass filter may be applied to reduce low-frequency components and external interference included in the signals. In one embodiment, MES candidate signals may be separated by retaining only signals having an amplitude higher than 1 to 6 dB compared to the background blood flow velocity. In one embodiment, MES candidate signals may be separated by retaining only signals having an amplitude higher than 3 dB compared to the background blood flow velocity. In one embodiment, MES candidate signals may be separated by filtering signals having an amplitude lower than 3 dB compared to the background blood flow velocity. In one embodiment, the first amplitude threshold value may be 1 to 6 dB. In one embodiment, the first amplitude threshold value may be 4 to 5 dB. In one embodiment, the first amplitude threshold value may be 3 dB.

[0086] In the step of additionally filtering (S240) a signal having a duration longer than a first duration threshold value among the filtered single or multiple signals, if the duration of the signal is maintained for a specific time (e.g., 0.5 to 1 second), the signal is likely to be a long-term blood flow change or noise and can therefore be removed. Conversely, short, transient signals are likely to be MES and can therefore be retained. In one embodiment, the first duration threshold value may be 0.5 to 2 seconds. In one embodiment, the first duration threshold value may be 0.5 to 1 second. In one embodiment, the first duration threshold value may be 100 to 1000% of the average signal duration of the MES. In one embodiment, the first duration threshold value may be 100 to 500%, 100 to 300%, or 100 to 200% of the average signal duration of the MES. In one embodiment, the average signal duration of the MES may be 0.5 to 1 second. In one embodiment, the average signal duration of the MES may be 0.75 seconds.

[0087] In one embodiment, a multi-frame comparison method may be applied during the signal filtering process, and reliability may be improved by assessing whether MES candidate signals occur repeatedly within a specific time interval. Furthermore, a compensation process may be added to ensure that key characteristics (e.g., amplitude, duration, frequency fluctuations) are maintained even after the signal is filtered. Finally, the filtered signal may be output in an analyzable form, and the analysis results may be stored. In an embodiment of the present invention, the preprocessed signal is stored in the processor and may be utilized as input data for subsequent artificial intelligence analysis.

[0088] In one embodiment, the steps of performing preprocessing of acquired data (S120) in the operating method of the MES detection device, continuously acquiring first Doppler ultrasound measurement results as a plurality of Doppler ultrasound images (S210), extracting single or multiple signals from the plurality of Doppler ultrasound images (S220), filtering signals having an amplitude lower than a first amplitude threshold value in the single or multiple signals, and additionally filtering signals having a duration longer than a first duration threshold value in the filtered single or multiple signals (S240) may each be independently used to perform preprocessing, or may be used in combination.

[0089] Figure 4 illustrates a learning method of an artificial intelligence algorithm used in an operating method of an MES detection device according to one embodiment of the present invention.

[0090] The first artificial intelligence algorithm may have been trained to match the preprocessed second Doppler ultrasound measurement results based on MES clinical data with the presence or absence of labeled MES.

[0091] In one embodiment, training data may be generated after preprocessing the raw signal measured by the probe. In one embodiment, the preprocessing process may be the same as the step of performing preprocessing of data acquired from the first Doppler ultrasound measurement results (S120). That is, the preprocessing process may include the steps of continuously acquiring the second Doppler ultrasound measurement results as a plurality of Doppler ultrasound images, extracting a single or multiple signals from the plurality of Doppler ultrasound images, filtering a signal having an amplitude lower than a second amplitude threshold from the single or multiple signals, and additionally filtering a signal having a duration longer than a second duration threshold from the filtered single or multiple signals. The second amplitude threshold may be the same as the first amplitude threshold. The second duration threshold may be the same as the first duration threshold. A signal obtained by preprocessing the Doppler ultrasound measurement results based on clinical data may be used as training data. The training data is divided into specific time intervals (e.g., 100 to 200 ms units) and may overlap to a certain extent to maintain signal continuity. Each signal can be normalized to emphasize the characteristics of the MES signal through frequency analysis and filtering to remove noise. Training data can include signal characteristics such as amplitude, frequency variation, and duration. Since the MES signal has a relatively high amplitude compared to the background blood flow, the data can be organized based on signal intensity. Using methods such as the Fourier transform, the signal variation in specific frequency bands can be analyzed. Since the MES signal occurs in a short period of time, it can be classified based on a specific duration. Training data can include Doppler ultrasound data collected in various environments, and the dataset can be classified based on the target site (head, neck, arm, etc.) and vessel type (middle cerebral artery, carotid artery, femoral artery, etc.).

[0092] In one embodiment, the labeling data may be correct data corresponding to the training data, and may include information indicating whether each Doppler ultrasound signal contains MES. In an embodiment of the present invention, signals may be labeled based on clinical data on the presence or absence of MES, enabling a machine learning model to learn the correct pattern. The labeling data may be constructed based on clinical research and analyses by medical experts, and may be classified to distinguish whether the Doppler ultrasound signal is an actual MES, normal blood flow, or artifact. Signals determined to be MES may be further grouped based on signal intensity, frequency pattern, and duration, and may be clearly distinguished from non-MES signals. To prevent data imbalance during neural network training, the labeled data may be variously transformed through an augmentation process. In an embodiment of the present invention, a labeling method such as binary classification or multi-class classification may be applied. In the case of binary classification, each Doppler ultrasound signal may be trained to distinguish whether it is MES or non-MES. For multi-classification, the MES signal can be trained to differentiate between different severity levels or to refine and distinguish specific signal patterns. MES clinical data can include Doppler ultrasound results collected in neurology and cardiology studies, patient data measured in hospitals, publicly available medical datasets, and self-built datasets. This data is labeled to enable a neural network model to distinguish between MES signals and normal blood flow signals, and can include feature vectors such as signal amplitude, frequency variation, and duration.

[0093] In one embodiment, the neural network model can utilize a Convolutional Neural Network (CNN) or a Long Short-Term Memory (LSTM). The CNN model can be used to learn features related to the MES and blood vessel location from preprocessed Doppler ultrasound image data. The LSTM model can be applied to learn signal change patterns over time and reflect signal continuity. A hybrid model combining CNN and LSTM can also be applied as needed. As a loss function, the cross-entropy loss function can be used during the model training process to evaluate learning performance. Algorithms such as Adam and RMSprop can be used as optimizers, and the learning rate can be adjusted to achieve optimal results. The trained model can be evaluated using a validation set and a test set. Model performance can be evaluated based on indicators such as accuracy, precision, recall, and F1 score.

[0094] FIG. 5 is a flowchart illustrating a method of operating an MES detection device according to one embodiment of the present invention for changing an amplitude threshold value and a duration threshold value as MES is detected.

[0095] The operating method of the MES detection device of the present invention may include a procedure of determining whether MES is detected (S310) after determining whether MES exists using a first artificial intelligence algorithm (S130) or providing information on microembolism within a target (S140), and if MES is detected, lowering a first amplitude threshold or raising a first duration threshold (S320), and if MES is not detected, maintaining the first amplitude threshold and the first duration threshold (S330).

[0096] In the step of determining whether MES is detected (S310), a process of analyzing a Doppler ultrasound signal to determine whether MES is detected may be performed. In an embodiment of the present invention, a first artificial intelligence algorithm performs signal analysis, and if MES is detected a certain number of times or more within a certain period of time, it may be determined that MES occurrence occurs in a continuous pattern. It may be analyzed whether the detected MES signal is a one-time event or a pattern that occurs continuously. If MES is repeatedly detected for a certain period of time, the amplitude and duration thresholds may be automatically adjusted to reflect the signal characteristics. Conversely, if MES is not detected for a certain period of time, the thresholds may be reset to the default values.

[0097] In the step (S320) of lowering the first amplitude threshold or raising the first duration threshold when an MES is detected, a process may be performed to determine whether to maintain or adjust the currently set amplitude threshold (first amplitude threshold) and duration threshold (first duration threshold). When a new MES pattern is detected, it may be evaluated whether the existing threshold is appropriate for optimizing MES detection. If the existing threshold is too high and may miss some MES signals, the amplitude threshold may be lowered or the duration threshold may be raised to increase detection sensitivity. Conversely, if necessary to reduce false positives (noise signals), the amplitude threshold may be increased or the duration threshold may be extended. In an embodiment of the present invention, the threshold change range may be dynamically adjusted, the amplitude threshold may be adjusted in the range of 1 to 6 dB, and the duration threshold may be adjusted in the range of 0.5 to 2 seconds.

[0098] The adjusted threshold value is stored in storage (140) and can be newly reflected in subsequent analysis processes. The updated threshold value can be applied to subsequent collected Doppler ultrasound signals to optimize detection sensitivity. A function to reset the threshold value to its default value if MES is not detected for a certain period of time may also be included. In embodiments of the present invention, the threshold value is dynamically optimized based on previous analysis results, thereby maintaining the accuracy of MES detection while flexibly responding to signal fluctuations.

[0099] FIG. 6 is a flowchart showing an operation method of an MES detection device having a blood vessel location recognition function according to one embodiment of the present invention.

[0100] Another method of operating an MES detection device of the present invention includes a step of obtaining a third Doppler ultrasound measurement result through a sensor module (S410), a step of performing preprocessing of the obtained data (S420), a step of obtaining a location of a blood vessel to be measured using a third artificial intelligence algorithm (S430), a step of obtaining a first Doppler ultrasound measurement result through the sensor module (S440), a step of performing preprocessing of the obtained data (S450), a step of determining whether there is an MES using the first artificial intelligence algorithm (S460), and a step of providing information on a microembolism in the object (S470).

[0101] In the step of acquiring the third Doppler ultrasound measurement result through the sensor module (S410), a process of acquiring a signal for blood vessel exploration may be performed using a probe. In one embodiment, the third probe may analyze the blood flow pattern to explore the location of the blood vessel to be measured. Since the blood vessel structure varies from person to person, it is necessary to precisely explore the blood vessel to be measured. The probe transmits a multi-frequency signal to measure the speed and direction of blood flow, and if a certain flow pattern is detected in a specific blood vessel, it can be set as the explored blood vessel location. In this step, data may be accumulated for a certain period of time to maintain the continuity of the blood flow signal, and may be utilized in a subsequent analysis process.

[0102] In one embodiment, the Doppler ultrasound analysis results may be provided through spectral Doppler ultrasonography. Spectral Doppler ultrasonography is a method for evaluating blood flow velocity by graphically measuring it and then quantifying it. The analysis results of spectral Doppler ultrasound may also be provided in the form of a Doppler signal waveform. In one embodiment, based on the acquired waveform data, the pattern of blood flow velocity changes can be analyzed and the location where a specific flow pattern remains constant can be searched for to determine the location of a blood vessel.

[0103] In one embodiment, the Doppler ultrasound analysis results may be provided as color Doppler ultrasonography. Color Doppler ultrasonography is a Doppler examination method that assigns colors to indicate the direction of blood flow and determines the brightness of the colors according to the speed of blood flow. Color Doppler ultrasonography can also provide images of the space being measured. The probe can visually represent the relative positions of blood vessels within a specific space based on blood flow, and can perform the function of exploring blood vessel structures in 2D or 3D space based on the ultrasound probe.

[0104] In one embodiment, the third Doppler ultrasound measurement result may be obtained from a tone Doppler ultrasound examination result, a color Doppler ultrasound examination result, or a combination of a tone Doppler ultrasound examination result and a color Doppler ultrasound examination result.

[0105] In the step of performing preprocessing of the acquired data (S420), the acquired third Doppler ultrasound measurement results may be preprocessed to prepare for blood vessel location analysis. In an embodiment of the present invention, various signal processing and normalization techniques may be applied to increase the accuracy of blood vessel location detection, and the goal is to maintain data quality and improve analyzability. In one embodiment, the process of preprocessing the third Doppler ultrasound measurement results may be the same as the process of preprocessing the first Doppler ultrasound measurement results.

[0106] In one embodiment, the third Doppler ultrasound measurement results may be provided as image data visually representing the direction and velocity of blood flow, and various preprocessing steps may be performed to analyze the data. Preprocessing of the color Doppler ultrasound image data may include steps such as image quality enhancement, color correction, spatial alignment, and feature extraction, which can more accurately analyze blood vessel locations and optimize signal processing for MES detection.

[0107] Color Doppler ultrasound scan results can contain various noises in the medical imaging environment, so filtering and image correction processes can be performed to improve signal quality. Gaussian filters or median filters can remove high-frequency noise that may occur in ultrasound images and emphasize continuous blood flow patterns. Binary thresholding and contrast enhancement can improve analysis accuracy by clearly distinguishing blood flow from surrounding tissue. Spatial smoothing techniques can minimize unnecessary background noise while maintaining the uniformity of blood flow.

[0108] Since color Doppler ultrasound results are color-coded according to the direction of blood flow and have brightness that varies with velocity, color correction can be performed to facilitate analysis. "RGB → HSV color space conversion" can convert the RGB (Red-Green-Blue) color space to the HSV (Hue-Saturation-Value) color space to more effectively analyze the color of ultrasound images. Color mapping analysis can highlight blood flow patterns of specific colors, allowing for more clear recognition of the flow in specific blood vessels. Channel separation can separate and analyze individual color channels to analyze blood flow direction based on color information.

[0109] Color Doppler ultrasound images can be spatially distorted depending on probe movement and the ultrasound measurement angle, so alignment and correction processes can be performed for analysis. Geometric transformation can correct image distortion caused by probe position changes. Region of Interest (ROI) setting allows analysis of specific blood vessels by setting a region of interest (ROI). Multi-frame alignment tracks blood vessel locations based on continuous image data and compensates for image loss in specific frames.

[0110] Preprocessed color Doppler ultrasound examination results can be analyzed to characterize blood flow and extract features that can be utilized for MES detection. Blood flow velocity-based feature analysis can quantify blood flow velocity based on changes in intensity and hue in color Doppler images, and by analyzing blood flow velocity change patterns, abnormal turbulence or vascular stenosis can be detected. Vascular structure-based feature analysis can recognize the shape of blood vessels in images, extract vessel diameters and directionality, detect continuous vessel patterns in color Doppler images, and optimize the location of specific vessels. Blood flow direction analysis and MES detection can assess the possibility of MES and classify the characteristics of signals when blood flow direction changes abruptly or multiple signal patterns are detected within a specific vessel.

[0111] The preprocessed color Doppler ultrasound examination results can be stored in storage (140) and subsequently utilized in the vessel location search and MES detection algorithm.

[0112] When the results of a transduced Doppler ultrasound examination (waveform data) and a color Doppler ultrasound examination (image data) are mixed, data alignment can be performed to facilitate analysis. Signal intensity can be normalized based on changes in blood flow velocity and frequency to minimize variations that may occur depending on the ultrasound measurement environment. For image-based data (color Doppler ultrasound results), color correction and sharpening can be applied, while for waveform data (transduced Doppler ultrasound results), high- and low-frequency components can be corrected to minimize noise.

[0113] Filtering techniques can be applied to remove external interference and environmental noise that may occur during the measurement process. High-pass filtering and low-pass filtering can be applied to improve signal quality. Optimizing the signal-to-noise ratio (SNR) in transducer Doppler ultrasound signals can retain only vascular signals and remove background signals.

[0114] If some signals are lost or data intervals are inconsistent in certain sections, interpolation techniques can be applied to supplement the data. Based on blood flow velocity change patterns, signals are adjusted to maintain a consistent flow, allowing the relative positions of blood vessels to be more clearly identified. Multi-frame analysis can be performed to supplement vascular information missed in a single frame with data from subsequent frames.

[0115] In this process, color Doppler ultrasound image data and tone Doppler ultrasound waveform data are combined to enable comprehensive analysis of vascular structure and blood flow velocity patterns.

[0116] In the step of obtaining the location of the target blood vessel using the third artificial intelligence algorithm (S430), a process of obtaining the final location of the target blood vessel using the third artificial intelligence algorithm can be performed based on the preprocessed blood vessel location candidate data. In an embodiment of the present invention, analysis can be performed using a neural network model by considering blood flow velocity, direction, flow pattern, etc., and it can be determined whether a specific blood vessel is an optimized target for MES detection.

[0117] In an embodiment of the present invention, the third artificial intelligence algorithm can apply a learned neural network model for Doppler ultrasound-based blood vessel location search.

[0118] Training data may include transient Doppler ultrasound examination results and color Doppler ultrasound examination results collected based on clinical data. In one embodiment, fourth-order Doppler ultrasound measurement results may be used as training data. The dataset may include location information of target blood vessels (such as the middle cerebral artery, carotid artery, and femoral artery) and blood flow signal patterns collected from those blood vessels. By including blood vessel data measured from various body parts, the dataset enables accurate blood vessel detection in specific areas.

[0119] Image analysis based on CNNs or LSTMs can be considered for analysis. CNNs can learn from color Doppler ultrasound data and automatically recognize vascular structures. LSTM models can be applied to analyze transient Doppler ultrasound waveform data and learn continuous changes in blood flow velocity. By combining the results of these two analyses, a neural network model can determine whether a specific blood vessel is suitable for MES detection.

[0120] The labeling data may include the optimal MES measurement vessel locations based on cerebral blood flow anatomy data. The relative positions of the vessels may be determined based on the probe, and various coordinate system representation methods such as Cartesian coordinates, polar coordinates, and cylindrical coordinates may be applied to represent the relative positions. In one embodiment, the cerebral blood flow anatomy data may be used to optimize vessel location analysis and MES detection using Doppler ultrasound, and may include information on major cerebral blood vessel structures and blood flow patterns. The data may be organized based on anatomical characteristics such as the anatomical arrangement of cerebral blood vessels, blood flow velocity and direction, and vessel diameter, thereby enabling accurate navigation of specific vessels and assessment of blood flow status. In one embodiment, the cerebral blood flow anatomy data may include information on major vessels such as the middle cerebral artery (MCA), anterior cerebral artery (ACA), posterior cerebral artery (PCA), basilar artery (BA), vertebral artery (VA), and carotid arteries (ICA, CCA, ECA). These data can include data collected across a range of age groups and pathological conditions (e.g., stroke, vascular stenosis, etc.) to reflect the anatomical variations of individual patients, and can improve the reliability of Doppler ultrasound analysis by providing a range of normal blood flow velocities for each vessel.

[0121] After determining the vessel location, a third AI algorithm can select the optimal target vessel for measurement by considering signal strength and flow patterns. To select a vessel suitable for MES detection, the blood flow intensity and pattern within the target vessel can be evaluated. If the blood flow within the vessel is too low or irregular, a different vessel can be selected to increase measurement reliability. Vessel locations can be updated in real time and automatically compensated for changes in the measurement environment.

[0122] The acquired vessel location data and third artificial intelligence algorithm can be stored in storage (140) and referenced during subsequent MES detection processes. By storing the relative locations of the vessels, an optimal measurement environment can be maintained when performing repeated tests at the same measurement site. Subsequently, when MES detection is performed, the probe can be automatically adjusted based on the detected vessel location.

[0123] The step of obtaining the first Doppler ultrasound measurement result through the sensor module (S440), the step of performing preprocessing of the obtained data (S450), the step of determining whether there is an MES using the first artificial intelligence algorithm (S460), and the step of providing information on microembolism within the subject (S470) may be the same as steps S110 to S140 of Fig. 1. In the step of providing information on microembolism within the subject (S470), the information on microembolism may include whether there is a microembolism, the location where the microembolism is found, the detection frequency of the microembolism, the signal intensity of the MES, or a combination thereof.

[0124] The embodiments of the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded on a computer-readable recording medium. In one embodiment, a non-transitory computer-readable recording medium that records a computer program for executing an MES detection method may be provided. The computer-readable recording medium may include program instructions, data files, data structures, etc., either singly or in combination. The program instructions recorded on the computer-readable recording medium may be specially designed and configured for the present invention or may be known and available to those skilled in the art of computer software. Examples of computer-readable recording 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 hardware devices specifically configured to store and execute program instructions, such as ROMs, RAMs, and flash memories. Examples of program instructions include not only machine language codes generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter, etc. A hardware device may be modified with one or more software modules to perform processing according to the present invention, and vice versa.

[0125] Although the present invention has been described above with specific details such as specific components and limited examples and drawings, these are provided only to help a more general understanding of the present invention, and the present invention is not limited to the above examples, and those with ordinary knowledge in the technical field to which the present invention pertains can make various modifications and changes based on this description.

[0126] Therefore, the idea of ​​the present invention should not be limited to the embodiments described above, and not only the scope of the patent claims described below but also all scopes equivalent to or equivalently modified from the scope of the patent claims are considered to fall within the scope of the idea of ​​the present invention.

[0127] [Explanation of symbols]

[0128] 100: MES detection device

[0129] 110: Sensor module

[0130] 112: Probe 1

[0131] 114: Second Probe

[0132] 120: Processor

[0133] 122: CPU

[0134] 124: NPU

[0135] 130: Motion Control Module

[0136] 140: Storage

[0137] 142: The First Artificial Intelligence Algorithm

[0138] 144: The Third Artificial Intelligence Algorithm

[0139] 146: First amplitude threshold

[0140] 148: First duration threshold

Claims

1. A method for providing information on whether a microembolism exists within a subject using Doppler ultrasound, A step of obtaining a first Doppler ultrasound measurement result by using a first probe on a measurement target part of the object; A step of preprocessing the first Doppler ultrasound measurement result by a processor to obtain a preprocessed first Doppler ultrasound measurement result; A step of determining whether a microembolic signal (MES) exists from the preprocessed first Doppler ultrasound measurement result using a first artificial intelligence algorithm by the above processor; and A step of providing information about microembolism within a subject based on the presence or absence of the MES by the processor; How to include.

2. A method according to claim 1, characterized in that the method of providing information on whether a microembolism exists in the object is performed in real time.

3. A method according to claim 1, characterized in that the measurement target area is the head, neck, arm, wrist, chest, thigh, or ankle.

4. In the first paragraph, the step of preprocessing the first Doppler ultrasound measurement result and obtaining the preprocessed first Doppler ultrasound measurement result is A step of continuously acquiring the first Doppler ultrasound measurement results as a plurality of Doppler ultrasound images by the processor; and A step of obtaining a first Doppler ultrasound measurement result preprocessed from the plurality of Doppler ultrasound images by the processor; Includes, A method characterized in that the plurality of Doppler ultrasound images are overlapped with each other so that no Doppler ultrasound signal is missing between the plurality of Doppler ultrasound images.

5. In the fourth paragraph, the step of obtaining a first Doppler ultrasound measurement result preprocessed from the plurality of Doppler ultrasound images is A step of extracting a single or multiple signals from the multiple Doppler ultrasound images by the above processor; A step of filtering a signal having an amplitude lower than a first amplitude threshold value from the single or multiple signals by the processor; and A step of obtaining a first Doppler ultrasound measurement result preprocessed from a single or multiple signals from which a signal having an amplitude lower than a first amplitude threshold value is filtered by the above processor. A method characterized by comprising:

6. In the fifth paragraph, the step of obtaining a first Doppler ultrasound measurement result preprocessed from a single or multiple signals from which a signal having an amplitude lower than the first amplitude threshold value is filtered A step of additionally filtering, by the processor, a signal having a duration longer than a first duration threshold from the filtered single or multiple signals; and A step of obtaining a preprocessed first Doppler ultrasound measurement result from a single or multiple signals from which signals having a duration longer than a first duration threshold are additionally filtered by the processor. A method characterized by comprising:

7. A method according to claim 1, wherein the first artificial intelligence algorithm is learned to match the second Doppler ultrasound measurement result preprocessed based on MES clinical data with the presence or absence of a labeled MES.

8. In the 6th paragraph, if the presence of MES is determined from the first Doppler ultrasound measurement result using the first artificial intelligence algorithm By the processor, lowering the first amplitude threshold, raising the first duration threshold, or lowering the first amplitude threshold and raising the first duration threshold, A method characterized in that, in the step of preprocessing the first Doppler ultrasound measurement result performed thereafter to obtain a preprocessed first Doppler ultrasound measurement result, a changed first amplitude threshold value, a changed first duration threshold value, or a changed first amplitude threshold value and a changed first duration threshold value are applied.

9. A method according to claim 1, characterized in that information on microembolism within the subject can be used to predict cerebral infarction, myocardial infarction, or cancer metastasis.

10. A method according to claim 1, characterized in that the method for providing information on whether a microembolism exists in the subject is performed by a wearable device or a surgical device.

11. In the first paragraph, the step of obtaining the first Doppler ultrasound measurement result is characterized in that the first Doppler ultrasound measurement result is obtained by using the first probe toward the measurement target blood vessel location of the measurement target area.

12. A method according to claim 11, characterized in that the blood vessel to be measured is a middle cerebral artery, internal carotid artery, external carotid artery, common carotid artery, brachial artery, brachiocephalic artery, internal thoracic artery, axillary artery, femoral artery, or femoral artery.

13. In the 11th paragraph, the location of the blood vessel to be measured is A step of obtaining a third Doppler ultrasound measurement result by using a third probe on a measurement target part of the object; A step of preprocessing the third Doppler ultrasound measurement result by the processor to obtain a preprocessed third Doppler ultrasound measurement result; and A step of obtaining the location of the blood vessel to be measured from the preprocessed third Doppler ultrasound measurement result using the third artificial intelligence algorithm by the above processor. A method characterized in that it is obtained through .

14. A method according to claim 11, characterized in that the measurement target blood vessel location is a relative location from the first probe or the third probe.

15. A method according to claim 14, characterized in that the first probe is directed toward the location of the blood vessel to be measured by adjusting the position and angle of the ultrasound probe of the first probe by the motion control module.

16. In the 14th paragraph, the third artificial intelligence algorithm is characterized in that it learns to match the labeled blood vessel location with the fourth Doppler ultrasound measurement result preprocessed based on cerebral blood flow anatomy data.

17. A method according to claim 14, characterized in that the information regarding microembolism in the subject includes the presence or absence of microembolism, the location where microembolism is detected, the detection frequency of microembolism, the signal intensity of MES, or a combination thereof.

18. A non-transitory computer-readable recording medium recording a computer program for executing the method according to paragraph 1.

19. A sensor module including a first probe; Storage for storing the first artificial intelligence algorithm; and A processor that uses the first probe to measure a target area of ​​the subject to obtain a first Doppler ultrasound measurement result, preprocesses the first Doppler ultrasound measurement result to obtain a preprocessed first Doppler ultrasound measurement result, determines whether MES exists from the preprocessed first Doppler ultrasound measurement result using the first artificial intelligence algorithm, and provides information on microembolism in the subject based on the presence or absence of the MES. MES detection device including.

Images