Smartphone-based carotid artery stenosis information provision method and system

Abstract

According to the present invention, a user can non-invasively identify a carotid artery stenosis state by using only a smartphone terminal without separate medical equipment or specialized ultrasonic diagnostic technology. Accordingly, the present invention can recognize the risk of stenosis or blood flow abnormality early even in an environment with low medical accessibility, and can be utilized for daily self-health management.

Description

Smartphone-based carotid artery stenosis information provision method and system thereof

[0001] The present invention relates to a field of technology that provides information on carotid artery stenosis by analyzing a user's carotid artery auscultation signal using artificial intelligence via a smartphone.

[0002] In general, carotid artery stenosis is one of the major causes of cardiovascular diseases such as stroke, so early diagnosis and continuous management are very important. Conventional diagnosis of carotid artery stenosis is performed using medical equipment such as ultrasound Doppler, CT, and MRI, but these devices have limitations as they are expensive and require specialized personnel.

[0003] With the recent advancements in smartphones and artificial intelligence (AI) enabling non-invasive biosignal analysis, there is a growing demand for technologies that allow users to check their vascular health status themselves. However, existing technologies are limited to measuring general biosignals such as heart rate, blood pressure, and oxygen saturation, making them insufficient for precisely assessing blood flow abnormalities like carotid artery stenosis.

[0004] Accordingly, there is a need for technology that can easily monitor a user's vascular health by collecting auscultation signals from the carotid artery area using a smartphone without a separate medical device, and analyzing and providing information on stenosis or blood flow abnormalities through artificial intelligence.

[0005] The objective of the present invention is to provide a non-invasive and highly accessible technology for providing information on carotid artery stenosis by allowing a user to acquire auscultation signals from the carotid artery area using only a smartphone without separate medical equipment or professional knowledge, and to analyze and provide information on whether there is carotid artery stenosis and blood flow abnormalities through an artificial intelligence model.

[0006] In addition, another objective of the present invention is to provide a function that guides the user to an appropriate measurement location where carotid sounds are optimally collected, using an artificial intelligence-based quality evaluation model, in order to solve the problem that the quality of auscultation signals may vary depending on the location of the smartphone terminal.

[0007] Furthermore, the present invention aims to implement a technology that can more reliably evaluate the degree of carotid artery stenosis by minimizing errors caused by physical differences between individuals and measurement environments through the training of an artificial intelligence model using auscultation data obtained by a specialist.

[0008] The present invention provides information regarding carotid artery stenosis by acquiring an arterial auscultation signal from the user's carotid artery using a microphone built into a smartphone terminal and analyzing the auscultation signal through an artificial intelligence (AI) model.

[0009] The above artificial intelligence model (first artificial intelligence model) is based on YAMNet, a neural network model used for acoustic signal classification, and is trained using training data generated by preprocessing bruit data. Through this, it learns the patterns of blood flow signals in which stenosis exists and calculates a stenosis score, risk, or probability index from the input auscultation signal.

[0010] Furthermore, prior to the analysis of the first artificial intelligence model, the present invention is configured to determine whether the user's smartphone terminal is placed at a suitable measurement location of the carotid artery using a second artificial intelligence model, and to guide the user to move the location through a screen, vibration, or sound output module. The second artificial intelligence model is also based on YAMNet and is trained using training data that includes data obtained by a specialist placing the terminal at a suitable measurement location of the carotid artery and auscultating.

[0011] Therefore, the present invention enables a user to acquire carotid artery auscultation signals using only a smartphone and intuitively check stenosis information through artificial intelligence analysis, and enables high-accuracy blood flow analysis by including auscultation quality evaluation and location guidance functions.

[0012] According to the present invention, a user can non-invasively check for carotid artery stenosis using only a smartphone terminal without the need for separate medical equipment or professional ultrasound diagnostic technology. Accordingly, even in environments with low access to medical care, the risk of blood flow abnormalities or stenosis can be recognized early and utilized for daily self-health management.

[0013] In addition, the present invention provides higher accuracy and reliability than existing simple signal discrimination methods by applying a YAMNet-based artificial intelligence model to precisely analyze carotid artery stenosis information from bruit signals. In particular, by utilizing auscultation data obtained by specialists as training data, robust performance can be maintained even with individual anatomical differences or variations in the measurement environment.

[0014] In addition, the present invention can improve the quality of auscultation signals by using a second artificial intelligence model to guide the appropriate measurement location of a smartphone terminal in real time. This allows the user to easily find the most suitable location in the carotid artery area and improve the accuracy of stenosis assessment. Therefore, the present invention is a personalized blood flow health monitoring technology utilizing artificial intelligence, and has the effect of significantly contributing to the early detection and preventive management of carotid artery stenosis.

[0015] Figure 1 shows the overall configuration of a system that analyzes carotid artery auscultation data and user information using a smartphone and an artificial intelligence model, and provides carotid artery stenosis information.

[0016] The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be modified from one embodiment to another without departing from the spirit and scope of the invention. It should also be understood that the location or arrangement of individual components within each embodiment may be modified without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not meant to be limiting, and the scope of the invention should be understood to encompass the scope claimed by the claims and all equivalents thereof. Similar reference numerals in the drawings indicate identical or similar components across various aspects.

[0017] Terms such as "first," "second," etc., may be used to describe various components, but said components should not be limited by said terms. These terms are used solely for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be named the second component, and similarly, the second component may be named the first component. The term "and / or" includes a combination of a plurality of related described items or any of a plurality of related described items.

[0018] When it is stated that one component is "connected" or "connected" to another component, it should be understood that while it may be directly connected or connected to that other component, there may also be other components in between. On the other hand, when it is stated that one component is "directly connected" or "directly connected" to another component, it should be understood that there are no other components in between.

[0019] The terms used in this application are used merely to describe specific embodiments and are not intended to limit the invention. The singular expression includes the plural expression unless the context clearly indicates otherwise. In this application, terms such as "comprising" or "having" are intended to specify the presence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof.

[0020] In this regard, terms such as “about,” “substantially,” etc., used throughout the specification, are used to mean at or near the stated value when inherent manufacturing and material tolerances are presented in the said meaning, and are used to prevent unscrupulous infringers from unfairly exploiting the disclosure in which precise or absolute values ​​are mentioned to aid in understanding the invention. The terms “step” or “step of” used throughout the specification of the invention do not mean “step for”.

[0021] In this specification, the term "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Additionally, one unit may be realized using two or more hardware, and two or more units may be realized by one hardware.

[0022] Some of the operations or functions described herein as being performed by a terminal, device, or device may instead be performed by a server connected to said terminal, device, or device. Likewise, some of the operations or functions described as being performed by a server may also be performed by a terminal, device, or device connected to said server.

[0023] Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as generally understood by those skilled in the art to which the present invention pertains. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an ideal or overly formal sense unless explicitly defined in this application.

[0024] Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the attached drawings. In order to facilitate an overall understanding of the present invention, the same reference numerals are used for identical components in the drawings, and redundant descriptions of identical components are omitted.

[0025] One embodiment of the present invention provides a method for providing information regarding carotid artery stenosis, characterized by using a microphone built into a smartphone terminal to input an arterial auscultation signal acquired from a user's carotid artery area, analyzing the input arterial auscultation signal by a first artificial intelligence model, and providing information regarding carotid artery stenosis from the carotid artery auscultation signal.

[0026] In one embodiment of the present invention, before analyzing the arterial auscultation signal, the step of calculating the signal strength in short window units of 20 milliseconds (ms) and masking and removing the section exceeding a threshold value may be further included.

[0027] In one embodiment of the present invention, the arterial auscultation signal may be decomposed using wavelet transform, and may include a step of removing noise by eliminating noise components for each frequency band.

[0028] In one embodiment of the present invention, the first artificial intelligence model may be trained and inferred using YAMNet, a neural network model used for acoustic signal classification.

[0029] In one embodiment of the present invention, preprocessed bruit data may be utilized as training data for learning the first artificial intelligence model.

[0030] In one embodiment of the present invention, the information regarding the carotid artery stenosis may be expressed as a stenosis score, a stenosis risk, or a stenosis probability index.

[0031] In one embodiment of the present invention, prior to analysis by the first artificial intelligence model, the invention may further include a step of guiding the user to move the terminal to a location where the carotid artery auscultation signal is optimally collected using the screen, vibration, or sound output module of the smartphone terminal.

[0032] In one embodiment of the present invention, the guidance step is performed by a second artificial intelligence model that calculates a quality index of a carotid artery auscultation signal, and may guide the terminal location to a suitable measurement location when the quality index is above a threshold.

[0033] In one embodiment of the present invention, the second artificial intelligence model may be trained and inferred using YAMNet, a neural network model used for acoustic signal classification.

[0034] In one embodiment of the present invention, data obtained by a specialist placing a terminal at a suitable location for carotid artery measurement and auscultating may be utilized as training data for learning the second artificial intelligence model.

[0035] In one embodiment of the present invention, information regarding carotid artery stenosis may be provided in the form of a visual graph or score through a user terminal screen and stored together with the user's health history data and life log data, so as to be configured to periodically monitor changes in the risk of carotid artery stenosis.

[0036] A non-transitory computer-readable recording medium on which a computer program for executing a method according to one embodiment of the present invention is recorded may be provided.

[0037] A system for performing a method according to one embodiment of the present invention may be provided, comprising a microphone module for acquiring an arterial auscultation signal from a user's carotid artery, a second artificial intelligence quality evaluation module for evaluating the quality of the signal and guiding the location of the terminal, and an output module for displaying the result.

[0038] One embodiment of the present invention provides a carotid artery stenosis detection device characterized by comprising: a signal acquisition module that acquires an arterial auscultation signal from a user's carotid artery area; a stenosis analysis module that inputs the auscultation signal into a first artificial intelligence model to calculate carotid artery stenosis information; a quality evaluation module that uses a second artificial intelligence model to evaluate the quality of the signal and guide the terminal location to a suitable measurement location; and an output module that visually provides the result.

[0039] Figure 1 illustrates the overall configuration of a system that analyzes carotid artery auscultation data and user information using a smartphone and an artificial intelligence model, and provides carotid artery stenosis information. Figure 1 illustrates the configuration of a smartphone-based carotid artery stenosis information provision system according to an embodiment of the present invention. As shown in the figure, the user provides input data such as lifelog data, personal health information, and smoking and drinking habits through a smartphone and a wearable device, and inputs patient status information, such as disability status, through medical history data. Carotid artery auscultation data acquired using the smartphone's microphone is transmitted to an artificial intelligence (AI) analysis model along with data from medical staff, and features related to carotid artery stenosis are extracted from a YAMNet-based neural network model to calculate a cerebral blood flow health score and stenosis risk. The calculated results are provided in real-time through the smartphone application screen and are presented to the user in the form of monthly reports and customized brain health educational materials.

[0040] Overall configuration and operation flow of the system

[0041] A smartphone-based carotid artery stenosis information providing system according to one embodiment of the present invention is configured to analyze an auscultation signal acquired from a user's carotid artery area based on artificial intelligence (AI) to calculate stenosis information and provide it to the user through a smartphone application screen.

[0042] This system is broadly composed of a signal collection stage, a preprocessing stage, an artificial intelligence analysis stage, and a result provision stage.

[0043] First, a microphone module built into the smartphone collects arterial sounds generated when the user places the device in close contact with the carotid artery. The collected signals are transmitted in real time to a signal processing module inside the smartphone, and the signal processing module analyzes the intensity and frequency distribution of the acoustic signals to remove noise and normalize the data in a form that preserves features related to stenosis.

[0044] Subsequently, the normalized data is input into the first artificial intelligence model, and features (feature vectors) related to stenosis are extracted through the acoustic analysis neural network (YAMNet) inside the model.

[0045] The extracted features are compared with a pre-trained Bruit dataset to determine the stenosis score, risk, and blood flow abnormalities.

[0046] The analysis results are displayed on the user interface (UI) of the smartphone application through the results delivery module, and users can check the stenosis status in real time or track changes in their carotid artery health through monthly reports.

[0047] The system according to the present invention includes a processor and storage of a smartphone terminal in which each of the above components is executed. The processor includes a central processing unit (CPU) and a graphics processing unit (GPU) of the smartphone, and performs computation and control of software modules executed within the smartphone.

[0048] The CPU is responsible for signal acquisition, preprocessing, data input / output, and user interface control, while the GPU performs inference operations for the artificial intelligence model (YAMNet), acoustic feature vector extraction, and parallel computations during the training process. In this way, the CPU and GPU cooperate to execute the preprocessing module, the first AI model, the second AI model, and the result delivery module, enabling AI computations to be performed smoothly in a smartphone hardware environment.

[0049] Storage is connected to the processor and stores software modules, AI model parameters, Bruit signal data, user input data, analysis results, and more within the smartphone device. Storage may include volatile memory (RAM) and non-volatile memory (flash memory, SSD, etc.); model weights and training data are stored in non-volatile memory and retained even after a restart. Additionally, storage can store data in an encrypted form to protect user privacy, or it can be linked with a cloud server to remotely back up and manage data.

[0050] The software according to the present invention consists of program instructions configured to be executed on the processor, and the program can be run in an operating system (OS) environment of a smartphone.

[0051] That is, each functional module of the present invention (signal acquisition module, preprocessing module, first artificial intelligence analysis module, second artificial intelligence quality evaluation module, result provision module, etc.) is realized in hardware by a processor and storage, and operates in software by executing in the form of an instruction set on the processor. Therefore, the present invention can be implemented in a combined form of hardware and software, and can be operated in the same manner not only on smartphones but also on other electronic devices (tablets, smartwatches, etc.).

[0052] In addition, the system includes a second artificial intelligence model for quality evaluation. The second AI model evaluates the quality of auscultation signals in real time, and if it determines that the terminal is not placed in a suitable measurement position, it guides the user to adjust the terminal's position via a screen, vibration, or sound output module. This enables the user to always acquire high-quality carotid artery auscultation data from the optimal position.

[0053] Preprocessing steps for carotid artery auscultation data

[0054] The preprocessing step according to the present invention is a step for removing non-physiological noise and environmental noise within the carotid artery auscultation signal and for stably securing a characteristic signal related to stenosis.

[0055] The raw auscultation signal input from the smartphone microphone module is first divided into short window units of 20 milliseconds (ms). For each window, the signal strength (RMS; Root Mean Square) is calculated, and sections exceeding a preset threshold are masked and excluded from data analysis.

[0056] This process serves to eliminate abnormal single signals (impulsive noise) such as finger contact sounds, surrounding conversations, and sudden movements.

[0057] The signal remaining after masking is decomposed into multiple frequency bands through a wavelet transform.

[0058] By calculating a noise threshold for each band and applying the BayesShrink or VisuShrink technique, environmental low-frequency or high-frequency noise is effectively removed while preserving frequency components related to stenosis (approximately 80–600 Hz).

[0059] The preprocessed signal undergoes a normalization process and is converted into the form of input data for the first artificial intelligence model.

[0060] Through such a preprocessing process, the present invention can reliably secure medical-grade carotid artery auscultation data even in a limited sensor environment like a smartphone, and can improve the accuracy and reliability of stenosis analysis.

[0061] 1st AI Model (YAMNet)-based Stenosis Analysis Step

[0062] The first artificial intelligence model according to the present invention is configured based on YAMNet, a neural network model suitable for acoustic signal classification, and learns carotid artery stenosis patterns using bruit data.

[0063] The model utilizes pre-trained acoustic classification weights and is fine-tuned to receive a spectrogram of a carotid artery auscultation signal as input and output the presence and degree of stenosis.

[0064] The preprocessed auscultation signal is converted into the form of MFCC (Mel-Frequency Cepstral Coefficient) or Mel-Spectrogram and provided to the model input layer. The model extracts temporal-frequency features of the auscultation signal through a convolution layer and an embedding layer, and calculates the probability of stenosis occurrence as a probability value.

[0065] At this time, the training data includes both normal blood flow signals (True) and Bruy signals with stenosis (False), and the model learns the difference between them to output a stenosis score or risk index.

[0066] The analyzed results are temporarily stored in the local storage of the smartphone device, and if the user consents, they are linked to a cloud server and managed as long-term data. Through this, users can visually check changes in the stenosis index over time, and medical staff can refer to the data as needed to assist in diagnosis.

[0067] Second AI model-based quality evaluation and location guidance stage

[0068] The second artificial intelligence model according to the present invention evaluates the quality of the carotid artery auscultation signal in real time and determines whether the terminal is placed at a suitable measurement position in the carotid artery area.

[0069] This model is also composed of an acoustic signal classification neural network based on YAMNet, and is trained using both data obtained by specialists directly placing the terminal at suitable carotid artery measurement locations and auscultating at unsuitable locations.

[0070] The quality evaluation module calculates features such as SNR (Signal-to-Noise Ratio), spectral periodicity, and energy distribution from real-time auscultation signals, and calculates a quality score based on these values.

[0071] When the quality grade is above a threshold, the measurement signal is determined to be normal, and when it is below the threshold, instructions such as “move the terminal slightly upward” or “tilt it to the left” are provided to the user through the screen, vibration, or voice output module.

[0072] Through this process, users can quickly find the optimal measurement location without specialized knowledge and obtain accurate auscultation data that AI models can utilize.

[0073] Result provision and user interface configuration

[0074] The analysis results of the present invention are provided visually through the screen of a smartphone application.

[0075] The screen displays the stenosis score and risk grade calculated in real time, along with a graph-type cerebral blood flow health index and comparison results with previous measurement records.

[0076] Users receive a step-by-step evaluation such as “normal,” “caution,” or “high risk” based on the level of stenosis risk, and accordingly, personalized lifestyle guidance or messages recommending a visit to a medical institution are provided.

[0077] In addition, this system automatically generates monthly reports to summarize the user's carotid artery stenosis trends and provides brain health educational materials generated by AI based on measurement data.

[0078] Data is stored encrypted on cloud servers, and users' personal information is anonymized and managed.

[0079] Examples and Variants

[0080] Example 1: The system was implemented on a smartphone terminal (iOS-based), and carotid artery auscultation was performed on 20 adult users. The AI ​​model used a YAMNet-based network trained with data labeled by specialists, and the accuracy of determining the presence of stenosis was approximately 93%.

[0081] Example 2: To verify the quality evaluation performance of the second artificial intelligence model, a test was performed using 200 sets of auscultation data obtained from suitable / unsuitable locations, and the accuracy of the model's measurement location discrimination was confirmed to be 97%.

[0082] Example 3: As a result of comparing the performance of AI models with and without the wavelet preprocessing step, the model with the preprocessing applied showed approximately 12% higher accuracy in a noisy environment. This demonstrates that the preprocessing configuration of the present invention enables meaningful stenosis signal analysis even in a smartphone environment.

[0083] Variations: In addition to YAMNet, other neural network models such as CNN-LSTM, ResNet-Audio, and Whisper can be applied, and the system can be implemented in the form of smartwatches, earphones, or necklace-type devices. Furthermore, by linking with a hospital server, it is possible for medical staff to remotely check user measurement data and utilize it for diagnostic assistance.

[0084] The embodiments according to 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-transient computer-readable recording medium may be provided for recording a computer program for executing an MES detection method. The computer-readable recording medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the computer-readable recording medium may be those specifically designed and configured for the present invention or those 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 ROM, RAM, and flash memory. Examples of program instructions include machine code, such as that generated by a compiler, as well as high-level language code that can be executed by a computer using an interpreter, etc. A hardware device may be changed into one or more software modules to perform processing according to the present invention, and vice versa.

[0085] Although the present invention has been described above with reference to specific details such as specific components, limited embodiments, and drawings, this is provided only to aid in a more comprehensive understanding of the invention, and the invention is not limited to the above embodiments, and a person skilled in the art to which the invention belongs can make various modifications and changes from this description.

[0086] Accordingly, the scope of the present invention should not be limited to the embodiments described above, and all scopes equivalent to or equivalently modified from the claims set forth below, as well as the claims set forth below, shall be considered to fall within the scope of the concept of the present invention.

Claims

Using a microphone built into a smartphone terminal, an arterial auscultation signal acquired from the user's carotid artery is used as input, and Analyzing input arterial auscultation signals by a first artificial intelligence model to provide information regarding carotid artery stenosis from carotid artery auscultation signals. A method for providing information regarding carotid artery stenosis characterized by A method according to claim 1, further comprising the step of calculating the signal strength in short window units of 20 milliseconds (ms) and masking and removing the section exceeding a threshold value before analyzing the arterial auscultation signal. A method according to claim 1, characterized in that the arterial auscultation signal is decomposed using a wavelet transform, and noise is removed by eliminating noise components for each frequency band. A method according to claim 1, wherein the first artificial intelligence model is trained and inferred using YAMNet, a neural network model used for acoustic signal classification. A method according to claim 1, characterized by utilizing preprocessed bruit data as training data for learning the first artificial intelligence model. A method according to claim 1, characterized in that the information regarding carotid artery stenosis is expressed as a stenosis score, stenosis risk, or stenosis probability index. A method according to claim 1, further comprising the step of guiding a user to move the terminal to a location where carotid artery auscultation signals are optimally collected using a screen, vibration, or sound output module of the smartphone terminal prior to analysis by the first artificial intelligence model. A method according to claim 7, wherein the guidance step is performed by a second artificial intelligence model that calculates a quality index of a carotid artery auscultation signal, and when the quality index is above a threshold, the terminal location is guided to a suitable measurement location. A method according to claim 8, wherein the second artificial intelligence model is trained and inferred using YAMNet, a neural network model used for acoustic signal classification. A method according to claim 8, characterized in that it utilizes data obtained by a specialist placing a terminal at a suitable location for carotid artery measurement and auscultating, as training data for learning the second artificial intelligence model. A method according to claim 1, characterized in that information regarding carotid artery stenosis is provided in the form of a visual graph or score through a user terminal screen and is stored together with the user's health history data and life log data, so as to be configured to periodically monitor changes in the risk of carotid artery stenosis. A non-transitory computer-readable recording medium on which a computer program for executing the method according to claim 1 is recorded. A carotid artery stenosis information providing system for performing the method according to claim 8, characterized by comprising: a microphone module for acquiring an arterial auscultation signal from a user's carotid artery region; a second artificial intelligence quality evaluation module for evaluating the quality of the signal and guiding the location of the terminal; and an output module for displaying the result. As a device for providing carotid artery stenosis information, A signal acquisition module that acquires arterial auscultation signals from the user's carotid artery area, A stenosis analysis module that inputs the above auscultation signal into a first artificial intelligence model to calculate carotid artery stenosis information, A quality evaluation module using a second artificial intelligence model that evaluates the quality of the above signal and guides the terminal location to a suitable measurement location, and A carotid artery stenosis detection device characterized by including an output module that visually provides the above results.

Images