WO2022015010A1 - 음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 - Google Patents
음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 Download PDFInfo
- Publication number
- WO2022015010A1 WO2022015010A1 PCT/KR2021/008956 KR2021008956W WO2022015010A1 WO 2022015010 A1 WO2022015010 A1 WO 2022015010A1 KR 2021008956 W KR2021008956 W KR 2021008956W WO 2022015010 A1 WO2022015010 A1 WO 2022015010A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- onset
- signal
- spectrogram
- cough
- server
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/0002—Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network
- A61B5/0015—Remote monitoring of patients using telemetry, e.g. transmission of vital signals via a communication network characterised by features of the telemetry system
- A61B5/0022—Monitoring a patient using a global network, e.g. telephone networks, internet
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/08—Measuring devices for evaluating the respiratory organs
- A61B5/0823—Detecting or evaluating cough events
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/6813—Specially adapted to be attached to a specific body part
- A61B5/6823—Trunk, e.g., chest, back, abdomen, hip
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7203—Signal processing specially adapted for physiological signals or for diagnostic purposes for noise prevention, reduction or removal
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7253—Details of waveform analysis characterised by using transforms
- A61B5/7257—Details of waveform analysis characterised by using transforms using Fourier transforms
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7235—Details of waveform analysis
- A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
- A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
- A61B5/7271—Specific aspects of physiological measurement analysis
- A61B5/7275—Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B7/00—Instruments for auscultation
- A61B7/003—Detecting lung or respiration noise
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/10—Machine learning using kernel methods, e.g. support vector machines [SVM]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
- G06N20/20—Ensemble learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/088—Non-supervised learning, e.g. competitive learning
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L17/00—Speaker identification or verification techniques
- G10L17/26—Recognition of special voice characteristics, e.g. for use in lie detectors; Recognition of animal voices
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L25/00—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00
- G10L25/48—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use
- G10L25/51—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination
- G10L25/66—Speech or voice analysis techniques not restricted to a single one of groups G10L15/00 - G10L21/00 specially adapted for particular use for comparison or discrimination for extracting parameters related to health condition
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H30/00—ICT specially adapted for the handling or processing of medical images
- G16H30/40—ICT specially adapted for the handling or processing of medical images for processing medical images, e.g. editing
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/30—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for calculating health indices; for individual health risk assessment
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/70—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for mining of medical data, e.g. analysing previous cases of other patients
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B2562/00—Details of sensors; Constructional details of sensor housings or probes; Accessories for sensors
- A61B2562/02—Details of sensors specially adapted for in-vivo measurements
- A61B2562/0204—Acoustic sensors
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
- A61B5/683—Means for maintaining contact with the body
- A61B5/6831—Straps, bands or harnesses
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B5/00—Measuring for diagnostic purposes; Identification of persons
- A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
- A61B5/6887—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient mounted on external non-worn devices, e.g. non-medical devices
- A61B5/6898—Portable consumer electronic devices, e.g. music players, telephones, tablet computers
Definitions
- the embodiment relates to a method for counting coughs based on an acoustic signal, a server for performing the same, and a non-transitory computer-readable recording medium.
- the number of coughs is the most basic and important indicator that can objectively evaluate the severity of coughing, but the only way in which trained personnel can count the number of times while listening to the cough is the only one.
- a quantitative monitoring platform is needed in addition to cough analysis through 24-hour recording at the hospital. Frequency), intensity, and disturbance in life due to coughing through questionnaires/questions to diagnose the severity and provide treatment, or 2) check the severity of coughing in hospitals and clinical trials of global pharmaceutical companies.
- a sound recorder is attached to the patient to record sound for tens of minutes to 24 hours, and then trained personnel listen to the recording and count the number of coughs, so medical information with objectivity and efficiency is insufficient for both patients and medical staff.
- An embodiment has an object to provide a method for accurately counting a cough based on an acoustic signal.
- the onset signal includes a signal corresponding to an attack of a sound and having a predetermined length in the time domain; obtaining a spectrogram corresponding to the extracted onset signal; determining whether the acquired spectrogram is a cough section using a cough discrimination model; and calculating the total number of coughs in the acoustic signal based on the determined result of the spectrogram of the onset signal, wherein the acquiring and the determining include: the one or more extracted onsets It is performed for each onset signal of the signal, and in the step of calculating the total number of coughs, if the time point of the first onset signal and the time point of the second onset signal are spaced apart by more than a reference time, it is counted as two, and the first The method may be provided, wherein if the timing of the onset signal
- a communication unit for acquiring an acoustic signal including a sound recorded from an external device; a memory unit for storing instructions for loading the cough discrimination model; extracting one or more onset signals from the acoustic signal, wherein the onset signal includes a signal corresponding to an attack of a sound and has a predetermined length in a time domain; and a spectrometer corresponding to the extracted onset signal gram, determine whether the acquired spectrogram is a cough section by using the cough discrimination model, and calculate the total number of coughs in the acoustic signal based on the determined result of the spectrogram of the onset signal a control unit configured to do so;
- a server comprising a may be provided.
- a method of accurately counting coughs while reducing the amount of computation to be processed is provided.
- FIG. 1 is a diagram schematically showing the configuration of the entire system 10 for counting cough according to an embodiment of the present application.
- FIG. 2 is a diagram for explaining the components of the server 2000 according to an embodiment of the present application.
- FIG. 3 is a flowchart illustrating a method of counting cough according to an embodiment of the present application.
- FIG. 4 is a diagram for explaining a method of detecting an onset signal according to an embodiment of the present application.
- FIG 5 is a diagram for explaining a module for detecting an onset signal included in the control unit 2300 according to an embodiment of the present application.
- FIG. 6 is a diagram for explaining a cough determination operation according to an embodiment of the present application.
- FIG. 7 is a view for explaining a cough counting method according to an embodiment of the present application.
- the onset signal includes a signal corresponding to an attack of a sound and having a predetermined length in the time domain; obtaining a spectrogram corresponding to the extracted onset signal; determining whether the acquired spectrogram is a cough section using a cough discrimination model; and calculating the total number of coughs in the acoustic signal based on the determined result of the spectrogram of the onset signal, wherein the acquiring and the determining include: the one or more extracted onsets It is performed for each onset signal of the signal, and in the step of calculating the total number of coughs, if the time point of the first onset signal and the time point of the second onset signal are spaced apart by more than a reference time, it is counted as two, and the first The method may be provided, wherein if the timing of the onset signal
- the length in the time domain of the onset signal may be longer than the reference time, the method may be provided.
- the extracting of the onset signal may include: detecting an onset point in the sound signal; and extracting a signal corresponding to the time interval of the predetermined length from the detected onset point as a starting point; including a, a method may be provided.
- the obtaining of the spectrogram may include obtaining a spectrogram by transforming the extracted onset signal into a frequency domain, and may include a Fourier transform on the extracted onset signal.
- the step of obtaining the spectrogram may include extracting a spectrogram corresponding to the extracted onset signal from all spectrograms obtained by converting the sound signal into a frequency domain. .
- the cough discrimination model is a classification model that receives spectrogram data and is trained to classify it as cough or non-cough
- the spectrogram data is a spectrogram image having a predetermined length in a time domain
- the cough discrimination model is learned using a training data set including the spectrogram data and tagging information labeled in the spectrogram data, and the tagging information determines whether the spectrogram data includes a sound corresponding to a cough.
- a method may be provided, including information regarding whether or not
- the determining may include: performing at least one preprocessing of resizing, scaling, and RGB conversion on the obtained spectrogram; and applying the preprocessed spectrogram to the cough discrimination model to determine whether the spectrogram is a cough section.
- the calculating step may include, with respect to the onset signals determined as coughing, determining whether an interval between time points of two adjacent onset signals in the time domain is within the reference time. .
- a non-transitory computer-readable recording medium for recording a computer program for executing the method may be provided.
- a communication unit for acquiring an acoustic signal including a sound recorded from an external device; a memory unit for storing instructions for loading the cough discrimination model; extracting one or more onset signals from the acoustic signal, wherein the onset signal includes a signal corresponding to an attack of a sound and has a predetermined length in a time domain; and a spectrometer corresponding to the extracted onset signal gram, determine whether the acquired spectrogram is a cough section by using the cough discrimination model, and calculate the total number of coughs in the acoustic signal based on the determined result of the spectrogram of the onset signal a control unit configured to do so;
- a server comprising a may be provided.
- a server may be provided in which the controller is configured to determine whether an interval between time points of two adjacent onset signals in the time domain is within the reference time.
- the controller performs at least one preprocessing of resizing, scaling, and RGB conversion on the obtained spectrogram, and uses the preprocessed spectrogram with the cough discrimination model to convert the obtained spectrogram.
- a server may be provided, configured to determine if the spectrogram is a cough zone.
- the cough discrimination model is a classification model that receives spectrogram data and is trained to classify it as cough or non-cough
- the spectrogram data is a spectrogram image having a predetermined length in a time domain.
- the cough discrimination model is learned using a training data set including the spectrogram data and tagging information labeled in the spectrogram data, and the tagging information determines whether the spectrogram data includes a sound corresponding to a cough.
- a server may be provided, including information regarding whether or not
- FIG. 1 is a diagram schematically showing the configuration of the entire system 10 for counting cough according to an embodiment of the present application.
- the system 10 may include a device 1000 and a server 2000 .
- the components shown in FIG. 1 are not essential, and the system 10 may have more components or fewer components.
- the device 1000 may be connected to the server 2000 through a network to transmit/receive necessary data.
- the network includes a local area network (LAN), a metropolitan area network (MAN), a wide area network (WAN), Wi-Fi, Wi-Fi Direct, It may include, but is not limited to, LTE Direct, and/or Bluetooth.
- the device 1000 may be fixed to a part of a person's body and perform a function of recording a sound for a predetermined time.
- An external sound may be acquired and recorded for a predetermined length of time (eg, 24 hours) through an input device such as a microphone of the device 1000 , and an acoustic signal including the recorded sound is transmitted to the server 2000 through a network. can be transmitted.
- the device 1000 may have a microphone facing the person's face, and may be fixed to the back of the person or fixed to the arm.
- a belt, an armband, etc. which are commonly used may be used.
- the device 1000 is a digital device to which a function capable of transmitting data by connecting to the server 2000 is applied, and may be a digital device having a memory means and a microprocessor equipped with an arithmetic capability.
- the device 1000 may be a wearable device such as smart glasses, a smart watch, a smart band, a smart ring, a smart necklace, or the like, or a smart phone, a smart pad, a desktop computer, a notebook computer, a workstation, a PDA, a web pad, a mobile phone, etc. It can be a rather traditional device such as
- the server 2000 may perform a function of counting coughs by analyzing an acoustic signal.
- the server 2000 may calculate the total number of coughs of the acoustic signal by extracting one or more onset signals from the acoustic signal and applying input data obtained based on the extracted onset signal to the cough discrimination model.
- the server 2000 may omit the operation of extracting one or more onset signals to determine cough/non-coughing, and applying all of the acoustic signals to the cough determination model, An advantage of rapidly reducing the amount of required server computation can be derived.
- the server 2000 in calculating the total number of coughs, the server 2000 counts one or two based on the time interval between the time points of the onset signals determined as two adjacent coughs through a pre-stored algorithm, An advantage of enabling accurate calculation of the total number of coughs without the problem of counting one cough with a long sound can be derived.
- the device 1000 may further include an application program for performing the method.
- an application may exist in the form of a program module in the device 1000 .
- the characteristics of such a program module may be generally similar to the communication unit 2100 , the memory unit 2200 , and the control unit 2300 , which will be described later.
- at least a part of the application may be replaced with a hardware device or a firmware device capable of performing substantially the same or equivalent function as the application, if necessary.
- the server 2000 may include a communication unit 2100 , a memory unit 2200 , and a control unit 2300 .
- the components shown in FIG. 2 are not essential, and the server 2000 may have more or fewer components.
- each component of the server 2000 may be physically included in one server, or may be a distributed server distributed for each function, but is not limited thereto.
- At least some of the communication unit 2100 , the memory unit 2200 , and the control unit 2300 of the server 2000 may be program modules that communicate with an external system (not shown).
- These program modules may be included in the server 2000 in the form of an operating system, an application program module, and other program modules, and may be physically stored in various known storage devices. Also, these program modules may be stored in a remote storage device capable of communicating with the server 2000 . Meanwhile, these program modules include, but are not limited to, routines, subroutines, programs, objects, components, and data structures that perform specific tasks or execute specific abstract data types to be described later according to the present application.
- the communication unit 2100 may serve as the server 2000 to transmit/receive data to/from an external device (eg, the device 1000 ).
- the communication unit 2100 may include one or more modules that enable communication.
- the communication unit 2100 may include a module that enables communication with an external device through a wired method.
- the communication unit 2100 may include a module that enables communication with an external device through a wireless method.
- the communication unit 2100 may include a module for communicating with an external device through a wired method and a module for enabling the communication unit 2100 to communicate with an external device through a wireless method.
- the communication unit 2100 is a wired communication module that accesses the Internet through a LAN (Local Area Network), etc., and a mobile communication network through a mobile communication base station, such as LTE (Long Term Evolution) for transmitting and receiving data.
- Communication module a short-distance communication module using a WLAN (Wireless Local Area Network) communication method such as Wi-Fi or a WPAN (Wireless Personal Area Network) communication method such as Bluetooth or Zigbee
- WLAN Wireless Local Area Network
- WPAN Wireless Personal Area Network
- Bluetooth Zigbee
- GNSS Global Navigation Satellite System
- GPS Global Positioning System
- the communication unit 2100 may receive an acoustic signal through the device 1000 . According to an embodiment of the present application, the communication unit 2100 may analyze the acoustic signal received from the device 1000 and transmit the total number of coughs for the acoustic signal to the device 1000 .
- the memory unit 2200 may store various data and programs necessary for the server 2000 to operate. For example, in the memory unit 2200 , an operating program (OS) for driving the server 2000 for counting coughs based on an acoustic signal and a method for counting coughs are performed in the server 2000 in order to be performed. Various programs to be driven or used, and various data related to media to be referenced by these programs may be stored.
- OS operating program
- Various programs to be driven or used, and various data related to media to be referenced by these programs may be stored.
- the memory unit 2200 may temporarily or semi-permanently store data. Examples of the memory unit 2200 include a hard disk (HDD), a solid state drive (SSD), a flash memory (1400, flash memory), a read-only memory (ROM), and a random access (RAM). Memory) or cloud storage. In addition, the memory unit 2200 may build and store a database for storing data, and is not limited thereto, and may be implemented with various modules for storing data.
- HDD hard disk
- SSD solid state drive
- flash memory (1400, flash memory
- ROM read-only memory
- RAM random access
- Memory or cloud storage.
- the memory unit 2200 may build and store a database for storing data, and is not limited thereto, and may be implemented with various modules for storing data.
- the memory unit 2200 may store a cough determination model. According to another embodiment of the present application, the memory unit 2200 may store an instruction for loading the cough determination model.
- Cough discrimination models include supervised learning algorithms (e.g., Logistic Regression, Support Vector Machine (SVM), Random Forest, etc.), unsupervised learning algorithms, and Artificial Neural Networks. ; ANN) or a deep learning algorithm such as a Fully-Connected Network or Convolutional Neural Network (CNN).
- supervised learning algorithms e.g., Logistic Regression, Support Vector Machine (SVM), Random Forest, etc.
- unsupervised learning algorithms e.g., unsupervised learning algorithms, and Artificial Neural Networks. ; ANN
- ANN Artificial Neural Networks.
- CNN Fully-Connected Network or Convolutional Neural Network
- the cough discrimination model may be a classification model trained to receive spectrogram data and output a mark for whether it is a cough section.
- the cough discrimination model may be a classification model that receives spectrogram data and is trained to classify it into cough or non-cough.
- the spectrogram data may be a spectrogram image having a predetermined length in the time domain.
- the cough discrimination model may be learned using a training data set including the spectrogram data and tagging information labeled on the spectrogram data.
- the tagging information may include information on whether the spectrogram data includes a sound corresponding to a cough.
- the tagging information may be a value selected based on information previously performed by a medical professional.
- the labeling of whether a sound corresponds to a cough may be determined based on data previously performed by a medical professional.
- the process of learning the cough discrimination model according to an embodiment of the present application is performed by the controller 2300 included in the server 2000 or an entity separate from the server 2000 (eg, the server 2000). and a learning server, etc.)
- an onset detection module may be stored in the memory unit 2200 .
- Each module may exist in the server 2000 as a separate module that is functionally distinguished, or may exist in the server 2000 as a physically integrated module by one subject performing the roles of a plurality of functional modules. The function and operation of each module will be clearly understood by the cough counting method described below, and a detailed description thereof will be omitted.
- the controller 2300 may perform a function of generalizing and controlling the overall operation of the server 2000 .
- the controller 2300 may perform calculation and processing of various types of information and may control operations of components of the server 2000 .
- the controller 2300 may be implemented as a computer or a similar device according to hardware, software, or a combination thereof.
- the control unit 2300 may be provided in the form of an electronic circuit such as a CPU chip that performs a control function by processing electrical signals in hardware, and may be provided in the form of a program that drives the hardware control unit in software.
- the controller 2300 may extract one or more onset signals from the sound signal received through the communication unit 2100 .
- the controller 2300 may extract one or more onset signals from the acoustic signal by using the onset detection module and the onset signal extraction module.
- the onset signal includes a signal corresponding to attack of sound and may have a predetermined length in the time domain.
- the controller 2300 may acquire a spectrogram corresponding to the extracted onset signal.
- the controller 2300 may obtain a spectrogram corresponding to the extracted onset signal by using the spectrogram conversion module.
- the control unit 2300 obtains the entire spectrogram of the acoustic signal by using the spectrogram conversion module, and crops only a portion corresponding to the extracted onset signal from among the entire spectrogram, which corresponds to the extracted onset signal.
- a spectrogram can be obtained.
- the controller 2300 may determine whether the acquired spectrogram is a cough section.
- the controller 2300 may determine whether it is a cough section by applying the spectrogram-based input data to the cough discrimination model.
- the controller 2300 may calculate the total number of coughs in the sound signal based on the determined result of the spectrogram of the onset signal. In this case, the controller 2300 may obtain a spectrogram corresponding to the onset signal, and perform an operation of determining whether it is a cough section based on the obtained spectrogram as many as the number of extracted onset signals. In the step of calculating the total number of coughs, the controller 2300 may compare the timing of the onset signal determined as coughing.
- the controller 2300 counts it as two, and the time point of the first onset signal and the time point of the second onset signal are the reference time If it is within the range, it can be counted as one, and in this case, the first onset signal and the second onset signal are signals determined to be coughing.
- FIG. 3 is a flowchart illustrating a method of counting cough according to an embodiment of the present application.
- the server 2000 may extract the onset signal (S1000).
- the server 2000 may extract an onset signal based on the sound signal received through the communication unit 2100 .
- the server 2000 may extract the onset signal from the received sound signal or may extract the onset signal by performing noise filtering on the received sound signal.
- FIG. 4 is a diagram for explaining a method of detecting an onset signal according to an embodiment of the present application.
- the server 2000 may extract an Onset Point (OP) from a Sound Signal (SS).
- OP Onset Point
- SS Sound Signal
- the server 2000 determines whether it is an onset point OP every sampling time (eg, 0.01 s), and when it is determined as an onset point OP, a predetermined length of time (Onset length) from the corresponding point in the acoustic signal SS , OL), an onset signal (OS) may be obtained by cropping the signal in the time interval corresponding to the .
- the onset signal OS may include a signal corresponding to an attack of a sound.
- the onset signal OS may have a predetermined length OL in the time domain. In this case, the number of the acquired onset signals OS may be the same as the number of the determined onset points OP.
- the method of determining whether the onset point OP is the onset point OP may be in the form of selecting a point where the onset intensity obtained for each predetermined sampling time is greater than the reference intensity as the onset point OP.
- the onset intensity may be obtained based on the Mel-Spectrogram passed through the Mel-filter bank.
- the controller 2300 may divide the acquired sound signal SS into two or more frequency sections, and add values obtained by passing each of the divided signals through the Mel-filter Bank to obtain the onset strength. When the obtained onset intensity is greater than the reference intensity, the controller 2300 may determine that it is the onset point OP.
- the controller 2300 may determine that it is the onset point OP.
- the reference intensity may be a predetermined constant value.
- the reference intensity may be variably defined with reference to the entire sound signal SS or the onset intensity around a point to be determined.
- FIG 5 is a diagram for explaining a module for detecting an onset signal included in the control unit 2300 according to an embodiment of the present application.
- the onset detection module 2310 may perform an operation of detecting an onset point based on the received acoustic signal.
- the onset detection module 2310 may detect an onset point while scanning the entire sound signal at a predetermined interval.
- the onset detection module 2310 may detect an onset point in the sound signal, and the detected onset point may be plural.
- the onset signal extraction module 2320 may extract an onset signal from the onset point detected by the onset detection module 2310 .
- the onset signal extraction module 2320 may extract an onset signal for each detected onset point.
- the onset signal extraction module 2320 may extract the onset signal by cutting the signal for a predetermined length of time using the detected onset point as a starting point.
- the first onset signal, the second onset signal, and the Nth onset signal may be extracted from the acoustic signal through the operations of the onset detection module 2310 and the onset signal extraction module 2320 .
- the predetermined interval at which the onset detection module 2310 performs scanning may be smaller than the length of the onset signal, and in this case, the first onset signal and the second onset signal may overlap.
- the server 2000 may acquire a spectrogram ( S2000 ).
- the server 2000 may obtain a spectrogram based on the extracted onset signal.
- the server 2000 may extract a spectrogram corresponding to the onset signal extracted from all spectrograms obtained based on the sound signal.
- a method of converting a signal having an amplitude value in the time domain into a spectrogram having an amplitude value in the frequency domain may be generally performed according to a known method.
- a signal to be converted is divided into a predetermined time interval, the divided signal is decomposed into individual sine waves through fast Fourier transform, and the amplitude signal according to the decomposed frequency is displayed as a frequency over time while displaying the amplitude through color.
- a spectrogram can be converted into a form. It is not limited thereto, and 'spectrogram' in the present application may include all spectrograms in a conventional sense.
- the spectrogram in the present application may be a mel spectrogram in which a frequency is converted to a mel scale.
- a spectrogram may be obtained.
- the spectrogram acquisition module may acquire as many spectrograms as the number of onset signals acquired by the onset signal extraction module 2320 .
- the spectrogram acquisition module may perform an operation of converting the onset signal into a spectrogram as much as the number of onset signals acquired by the onset signal extraction module 2320 .
- the spectrogram acquisition module converts the acoustic signal into a full spectrogram, and as many as the number of onset signals acquired by the onset signal extraction module 2320 from the entire spectrogram to the spectrogram corresponding to the onset signal An operation for extracting a gram may be performed.
- the server 2000 may determine whether it is a cough section (S3000). The server 2000 may determine whether it is a cough section based on the spectrogram obtained in step S2000 (S3000).
- FIG. 6 is a diagram for explaining a cough determination operation according to an embodiment of the present application.
- the cough determination module 2330 may determine whether it is a cough section based on a spectrogram corresponding to the onset signal.
- the cough determination module 2330 may include a cough determination model trained to receive spectrogram data and classify it as cough or non-cough.
- the spectrogram data may be a spectrogram having a predetermined length in the time domain.
- the spectrogram data may be a spectrogram corresponding to the onset signal extracted in step S1000.
- pre-processing may be performed by the data pre-processing module 2340 before the spectrogram data is input to the cough determination module 2330 .
- the spectrogram data is data on which at least one of resizing, scaling, and gray to RGB conversion has been performed through the data preprocessing module 2340 on the spectrogram corresponding to the onset signal extracted in step S1000. can be As a specific example, resizing may be performed on the spectrogram through the data preprocessing module 2340 .
- the cough discrimination model may be learned using a training data set including the spectrogram data and tagging information labeled on the spectrogram data.
- the tagging information may include information on whether the spectrogram data includes a sound corresponding to a cough.
- a cough discrimination model was constructed using EfficientNet.
- a spectrogram having a time length of 0.5 s was used as input data, and a cough/nasal cough marker evaluated by a person by listening to a recorded sound as output data was reflected.
- input data one spectrogram is divided by frequency section and divided into a total of 512 input nodes.
- the spectrogram is resized to a size of 300*300, and scaling is performed by dividing by the maximum value. , Gray to RGB conversion was performed.
- the cough discrimination model was trained using a total of 350,000 training data, and the cough discrimination model showed accuracy of 0.84 and recall 0.93.
- a spectrogram having a time length of 0.5 s was used as input data, and a cough/nasal cough marker evaluated by a person by listening to a recorded sound as output data was reflected.
- input data one spectrogram is divided by frequency section and divided into a total of 512 input nodes.
- the spectrogram is resized to a size of 300*300, and scaling is performed by dividing by the maximum value. , Gray to RGB conversion was performed.
- the cough discrimination model was trained using a total of 350,000 training data, and at this time, the cough discrimination model showed accuracy of 0.92 with precision and 0.9 of recall.
- Efficient Net B5 is known to be more accurate, but in actual experiments, it was confirmed that Efficient Net B3 had better precision. confirmed that it can affect
- the server 2000 may calculate the total number of coughs ( S4000 ).
- the server 2000 may calculate the total number of coughs in the acoustic signal based on the determined result of the spectrogram of the at least one onset signal obtained in step S3000.
- steps S2000 and S3000 may be performed for each onset signal of one or more extracted onset signals. Steps S2000 and S3000 may be performed a number of times corresponding to the number of onset signals acquired in S1000.
- the server 2000 may determine the total number of coughs based on the onset signal determined as coughing in S3000.
- the server 2000 may determine the total number of coughs by counting the onset signals determined as coughing.
- the cough section can be captured from the entire sound signal. Therefore, this operation can be performed to reduce the amount of computation by capturing the cough section and applying it to the cough discrimination model.
- the server 2000 may be implemented to count as one when the interval time between the time points of two adjacent onset signals determined as coughing is within a reference time (eg, 0.5 s).
- FIG. 7 is a view for explaining a cough counting method according to an embodiment of the present application.
- the 'time point' in the left column indicates the time value of the time point of the onset signal in the acoustic signal. In other words, if the onset point is detected at 15.5 seconds from the start of the sound signal, the time point of the onset signal may appear as 00:00:15.5.
- the 'tag' in the right column indicates the result value determined by the cough discrimination model based on the onset signal.
- the tag of the onset signal may appear as 'cough'.
- the server 2000 may count as one cough when the interval time between the time points of two adjacent onset signals is within the reference time. That is, the first onset signal determined as coughing (00:00:15.5), the second onset signal determined as coughing (00:00:15.7), the third onset signal determined as coughing (00:17:20), In the fourth onset signal determined as cough (00:17:20.5) and the fifth onset signal determined as coughing (00:22:00), the server 2000 determines the timing of the first onset signal and the timing of the second onset signal When it is spaced apart by 0.2 sec and is less than the reference time (0.25 sec), it can be counted as one cough. When the time point of the third onset signal and the time point of the fourth onset signal are spaced apart by 0.5 sec and greater than the reference time (0.25 sec), the server 2000 may count as two coughs.
- the server 2000 may not acquire the time interval between the time point of the second onset signal and the time point of the third onset signal.
- the server 2000 may continue the cough counting procedure by comparing the time points of the third onset signal and the fourth onset signal without obtaining a separation time between the time point of the second onset signal and the time point of the third onset signal.
- the reference time may be less than 0.5sec. This may be because it is rare for a person to complete two coughs in 0.5 seconds. Therefore, according to an embodiment of the present application, the reference time may be set to 0.25 sec, and is not limited thereto, and if necessary (eg, when considering the characteristic of continuous cough), an adjusted value may be used. can be used According to an embodiment of the present application, the onset length may be greater than 0.5sec. This may be because people complete most coughs within 0.5 seconds, so 0.5 seconds is long enough to determine whether a cough is present or not.
- the onset length may be set to 0.5 sec, and is not limited thereto, and an adjusted value may be used if necessary (eg, when determining the type of cough). have.
- the length of the onset signal in the time domain ie, the onset length
- the above-described embodiments according to the present invention may be implemented in the form of program instructions that can be executed through various computer components and recorded in a non-transitory computer-readable recording medium.
- the non-transitory computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination.
- the program instructions recorded in the non-transitory computer-readable recording medium may be specially designed and configured for the present invention, or may be known and used by those skilled in the computer software field.
- non-transitory computer-readable recording medium examples 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 floppy disks ( magneto-optical media), and hardware devices specially configured to store and execute program instructions such as ROM, RAM, flash memory, and the like.
- program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.
- the hardware device may be configured to operate as one or more software modules for carrying out the processing according to the present invention, and vice versa.
Landscapes
- Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Physics & Mathematics (AREA)
- Medical Informatics (AREA)
- Public Health (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Molecular Biology (AREA)
- Pathology (AREA)
- Artificial Intelligence (AREA)
- Biophysics (AREA)
- Animal Behavior & Ethology (AREA)
- Surgery (AREA)
- Heart & Thoracic Surgery (AREA)
- Veterinary Medicine (AREA)
- Signal Processing (AREA)
- Physiology (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Pulmonology (AREA)
- Software Systems (AREA)
- Psychiatry (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Multimedia (AREA)
- Computational Linguistics (AREA)
- Databases & Information Systems (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Acoustics & Sound (AREA)
- Fuzzy Systems (AREA)
- Radiology & Medical Imaging (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
Abstract
Description
Claims (15)
- 음향 신호를 분석하여 기침을 계수하는 방법에 있어서,상기 음향 신호에서 하나 이상의 온셋 신호를 추출하는 단계-상기 온셋 신호는 소리의 어텍(Attack)에 대응되는 신호를 포함하고, 시간 도메인에서의 기결정된 길이를 가짐-;상기 추출된 온셋 신호에 대응되는 스펙트로그램을 획득하는 단계;기침 판별 모델을 이용하여, 상기 획득된 스펙트로그램이 기침 구간인지 판별하는 단계; 및상기 온셋 신호의 스펙트로그램에 대하여 판별된 결과에 기초하여, 상기 음향 신호에서의 전체 기침 횟수를 계산하는 단계;를 포함하고,상기 획득하는 단계 및 상기 판별하는 단계는, 상기 추출된 하나 이상의 온셋 신호의 각 온셋 신호에 대해서 수행되고,상기 전체 기침 횟수를 계산하는 단계에서는,제1 온셋 신호의 시점과 제2 온셋 신호의 시점이 기준 시간을 초과하는 만큼 이격되어 있으면 둘로 계수되고,상기 제1 온셋 신호의 시점과 상기 제2 온셋 신호의 시점이 기준 시간 이내이면 하나로 계수되며,상기 제1 온셋 신호 및 상기 제2 온셋 신호는 기침으로 판별된 신호인, 방법.
- 제1 항에 있어서,상기 온셋 신호의 시간 도메인에서의 길이는 상기 기준 시간보다 긴, 방법.
- 제1 항에 있어서,상기 온셋 신호를 추출하는 단계는,상기 음향 신호에서 온셋 지점을 검출하는 단계; 및상기 검출된 온셋 지점을 시점으로 상기 기결정된 길이의 시간 구간에 대응되는 신호를 추출하는 단계;를 포함하는, 방법.
- 제3 항에 있어서,상기 스펙트로그램을 획득하는 단계는,상기 추출된 온셋 신호를 주파수 도메인으로 변환하여 스펙트로그램을 획득하는 단계이고, 상기 추출된 온셋 신호에 대한 퓨리에 변환을 포함하는, 방법.
- 제3 항에 있어서,상기 스펙트로그램을 획득하는 단계는,상기 음향 신호를 주파수 도메인으로 변환하여 획득된 전체 스펙트로그램에서 상기 추출된 온셋 신호에 대응되는 스펙트로그램을 추출하는 단계를 포함하는, 방법.
- 제1 항에 있어서,상기 기침 판별 모델은 스펙트로그램 데이터를 입력받아 기침 또는 비기침으로 분류하도록 학습된 분류모델이고,상기 스펙트로그램 데이터는, 시간도메인에서 기 결정된 길이를 가지는 스펙트로그램 이미지인, 방법.
- 제6 항에 있어서,상기 기침 판별 모델은 상기 스펙트로그램 데이터 및 상기 스펙트로그램 데이터에 라벨링된 태깅 정보를 포함하는 학습 데이터 셋을 이용하여 학습되고,상기 태깅 정보는 상기 스펙트로그램 데이터가 기침에 대응되는 소리를 포함하는지 여부에 관한 정보를 포함하는, 방법.
- 제1 항에 있어서,상기 판별하는 단계는,상기 획득된 스펙트로그램에 대하여 리사이징(Resizing), 스케일링(Scaling) 및 RGB 변환 중 적어도 하나의 전처리를 수행하는 단계; 및전처리된 스펙트로그램을 상기 기침 판별 모델에 적용하여, 상기 스펙트로그램이 기침 구간인지 판별하는 단계;를 포함하는, 방법.
- 제1 항에 있어서,상기 계산하는 단계는,기침으로 판별된 온셋 신호 들에 대하여, 시간 도메인에서 인접한 두 온셋 신호의 시점 사이의 간격이 상기 기준 시간 이내인지 판단하는 단계;를 포함하는, 방법.
- 제1 항의 방법을 실행하기 위한 컴퓨터 프로그램을 기록하는 비일시성의 컴퓨터 판독 가능 기록 매체.
- 음향 신호를 분석하여 기침을 계수하는 시스템에 있어서,외부 장치에서 녹음된 소리를 포함하는 음향 신호를 획득하는 통신부;기침 판별 모델을 로딩하기 위한 인스트럭션을 저장하는 메모리부;상기 음향 신호에서 하나 이상의 온셋 신호를 추출하고-상기 온셋 신호는 소리의 어텍(Attack)에 대응되는 신호를 포함하고, 시간 도메인에서의 기결정된 길이를 가짐-,상기 추출된 온셋 신호에 대응되는 스펙트로그램을 획득하고,상기 기침 판별 모델을 이용하여, 상기 획득된 스펙트로그램이 기침 구간인지 판별하고,상기 온셋 신호의 스펙트로그램에 대하여 판별된 결과에 기초하여, 상기 음향 신호에서의 전체 기침 횟수를 계산하도록 구성되는 제어부; 를 포함하는, 서버.
- 제11 항에 있어서,상기 제어부는,기침으로 판별된 온셋 신호 들에 대하여, 시간 도메인에서 인접한 두 온셋 신호의 시점 사이의 간격이 상기 기준 시간 이내인지 판단하도록 구성되는, 서버.
- 제11 항에 있어서,상기 제어부는, 상기 획득된 스펙트로그램에 대하여 리사이징(Resizing), 스케일링(Scaling) 및 RGB 변환 중 적어도 하나의 전처리를 수행하고,전처리된 스펙트로그램을 상기 기침 판별 모델을 이용하여, 상기 획득된 스펙트로그램이 기침 구간인지 판별하도록 구성되는, 서버.
- 제11 항에 있어서,상기 기침 판별 모델은, 스펙트로그램 데이터를 입력받아 기침 또는 비기침으로 분류하도록 학습된 분류모델이고,상기 스펙트로그램 데이터는, 시간도메인에서 기 결정된 길이를 가지는 스펙트토그램 이미지인, 서버.
- 제14 항에 있어서,상기 기침 판별 모델은 상기 스펙트로그램 데이터 및 상기 스펙트로그램 데이터에 라벨링된 태깅 정보를 포함하는 학습 데이터 셋을 이용하여 학습되고,상기 태깅 정보는 상기 스펙트로그램 데이터가 기침에 대응되는 소리를 포함하는지 여부에 관한 정보를 포함하는, 서버.
Priority Applications (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR1020237001174A KR20230028370A (ko) | 2020-07-13 | 2021-07-13 | 음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 |
| US17/987,104 US11877841B2 (en) | 2020-07-13 | 2022-11-15 | Method for counting coughs by analyzing sound signal, server performing same, and non-transitory computer-readable recording medium |
| US18/544,814 US12138035B2 (en) | 2020-07-13 | 2023-12-19 | Method for counting coughs by analyzing sound signal, server performing same, and non-transitory computer-readable recording medium |
| US18/904,531 US20250017489A1 (en) | 2020-07-13 | 2024-10-02 | Method for counting coughs by analyzing sound signal, server performing same, and non-transitory computer-readable recording medium |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR10-2020-0086464 | 2020-07-13 | ||
| KR20200086464 | 2020-07-13 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US17/987,104 Continuation US11877841B2 (en) | 2020-07-13 | 2022-11-15 | Method for counting coughs by analyzing sound signal, server performing same, and non-transitory computer-readable recording medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2022015010A1 true WO2022015010A1 (ko) | 2022-01-20 |
Family
ID=79555612
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/KR2021/008956 Ceased WO2022015010A1 (ko) | 2020-07-13 | 2021-07-13 | 음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 |
Country Status (3)
| Country | Link |
|---|---|
| US (3) | US11877841B2 (ko) |
| KR (1) | KR20230028370A (ko) |
| WO (1) | WO2022015010A1 (ko) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117064330A (zh) * | 2022-12-13 | 2023-11-17 | 上海市肺科医院 | 一种声音信号处理方法及装置 |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20230368000A1 (en) * | 2022-05-11 | 2023-11-16 | Covid Cough, Inc. | Systems and methods for acoustic feature extraction and dual splitter model |
| US20250246200A1 (en) * | 2024-01-31 | 2025-07-31 | Iulian-Alexandru Circo | Methods for automatic cough detection and uses thereof |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101122590B1 (ko) * | 2011-06-22 | 2012-03-16 | (주)지앤넷 | 음성 데이터 분할에 의한 음성 인식 장치 및 방법 |
| US20120071777A1 (en) * | 2009-09-18 | 2012-03-22 | Macauslan Joel | Cough Analysis |
| US20140336537A1 (en) * | 2011-09-15 | 2014-11-13 | University Of Washington Through Its Center For Commercialization | Cough detecting methods and devices for detecting coughs |
| KR20140142330A (ko) * | 2012-03-29 | 2014-12-11 | 더 유니버서티 어브 퀸슬랜드 | 환자 소리들을 처리하기 위한 방법 및 장치 |
| US20200015709A1 (en) * | 2017-02-01 | 2020-01-16 | ResApp Health Limited | Methods and apparatus for cough detection in background noise environments |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6168568B1 (en) * | 1996-10-04 | 2001-01-02 | Karmel Medical Acoustic Technologies Ltd. | Phonopneumograph system |
| US6436057B1 (en) * | 1999-04-22 | 2002-08-20 | The United States Of America As Represented By The Department Of Health And Human Services, Centers For Disease Control And Prevention | Method and apparatus for cough sound analysis |
| US20210319804A1 (en) * | 2020-04-01 | 2021-10-14 | University Of Washington | Systems and methods using neural networks to identify producers of health sounds |
-
2021
- 2021-07-13 WO PCT/KR2021/008956 patent/WO2022015010A1/ko not_active Ceased
- 2021-07-13 KR KR1020237001174A patent/KR20230028370A/ko active Pending
-
2022
- 2022-11-15 US US17/987,104 patent/US11877841B2/en active Active
-
2023
- 2023-12-19 US US18/544,814 patent/US12138035B2/en active Active
-
2024
- 2024-10-02 US US18/904,531 patent/US20250017489A1/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20120071777A1 (en) * | 2009-09-18 | 2012-03-22 | Macauslan Joel | Cough Analysis |
| KR101122590B1 (ko) * | 2011-06-22 | 2012-03-16 | (주)지앤넷 | 음성 데이터 분할에 의한 음성 인식 장치 및 방법 |
| US20140336537A1 (en) * | 2011-09-15 | 2014-11-13 | University Of Washington Through Its Center For Commercialization | Cough detecting methods and devices for detecting coughs |
| KR20140142330A (ko) * | 2012-03-29 | 2014-12-11 | 더 유니버서티 어브 퀸슬랜드 | 환자 소리들을 처리하기 위한 방법 및 장치 |
| US20200015709A1 (en) * | 2017-02-01 | 2020-01-16 | ResApp Health Limited | Methods and apparatus for cough detection in background noise environments |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117064330A (zh) * | 2022-12-13 | 2023-11-17 | 上海市肺科医院 | 一种声音信号处理方法及装置 |
| CN117064330B (zh) * | 2022-12-13 | 2024-04-19 | 上海市肺科医院 | 一种声音信号处理方法及装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| US20250017489A1 (en) | 2025-01-16 |
| US12138035B2 (en) | 2024-11-12 |
| KR20230028370A (ko) | 2023-02-28 |
| US20230071233A1 (en) | 2023-03-09 |
| US11877841B2 (en) | 2024-01-23 |
| US20240115158A1 (en) | 2024-04-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2022015010A1 (ko) | 음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 | |
| WO2021015381A1 (en) | Pulmonary function estimation | |
| WO2015008936A1 (ko) | 습관을 이용한 진단 장치 및 진단 관리 장치와 그를 이용한 진단 방법 | |
| US20220104725A9 (en) | Screening of individuals for a respiratory disease using artificial intelligence | |
| WO2019190016A1 (ko) | 인지 기능 재활 훈련 방법 및 장치 | |
| CN111553266A (zh) | 识别验证方法、装置及电子设备 | |
| WO2019031794A1 (ko) | 피검체의 치명적 증상의 발생을 조기에 예측하기 위한 예측 결과를 생성하는 방법 및 이를 이용한 장치 | |
| WO2019235828A1 (ko) | 투 페이스 질병 진단 시스템 및 그 방법 | |
| WO2022019402A1 (ko) | 시계열 생체 신호 기반의 인공신경망 모델 학습 컴퓨터 프로그램 및 방법 | |
| WO2021085947A1 (en) | Parkinson's disease diagnostic application | |
| CN114420302A (zh) | 一种企事业单位智能防疫控制系统 | |
| WO2018105995A2 (ko) | 빅데이터를 활용한 건강정보 예측 장치 및 방법 | |
| WO2025058347A1 (ko) | 인공지능 기반의 생체 데이터 분석을 통한 패혈성 쇼크 조기 예측방법, 장치 및 컴퓨터프로그램 | |
| WO2017090815A1 (ko) | 관절 가동 범위를 측정하는 장치 및 방법 | |
| KR20230011054A (ko) | 음향 신호를 분석하여 기침을 계수하는 방법, 이를 수행하는 서버 및 비일시성의 컴퓨터 판독 가능 기록 매체 | |
| WO2023063772A1 (ko) | 딥 러닝을 이용한 이미지 분석 기반의 피부 진단 시스템 및 방법 | |
| WO2024162605A1 (ko) | 심잡음을 기반으로 심장 판막 이상을 분석하기 위한 방법, 컴퓨터 프로그램, 및 시스템 | |
| WO2016200243A1 (ko) | 미병 분류를 보조하는 컴퓨팅 장치 및 방법 | |
| WO2022119347A1 (ko) | 초음파 영상 기반의 딥 러닝을 통한 관상동맥 경화반 조직 분석 방법, 장치 및 기록매체 | |
| WO2021096279A1 (ko) | 내시경 검사 중 병변이 발견된 위치에서의 데이터 입력 방법 및 상기 데이터 입력 방법을 수행하는 컴퓨팅 장치 | |
| WO2022050459A1 (en) | Method, electronic device and system for generating record of telemedicine service | |
| WO2025033683A1 (ko) | 건강 검진 데이터를 이용한 디지털 헬스케어 콘텐츠 제공 시스템 및 그 방법 | |
| CN117481611A (zh) | 一种高准确性的双模态eeg自动睡眠分期方法及系统 | |
| WO2023080697A1 (ko) | 심장 신호 분할 방법 및 이를 이용한 심장 신호 분할용 디바이스 | |
| WO2019139447A1 (ko) | 생체 데이터를 이용한 컨텐츠 평가 장치 및 이를 이용한 컨텐츠 평가 방법 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21841462 Country of ref document: EP Kind code of ref document: A1 |
|
| ENP | Entry into the national phase |
Ref document number: 20237001174 Country of ref document: KR Kind code of ref document: A |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 12/05/2023) |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 21841462 Country of ref document: EP Kind code of ref document: A1 |