# US 2016/0354053 A1 · Method and system for recognizing physiological sound

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2016/0354053 A1
- Document ID: US20160354053A1
- Status: Published application
- Kind: A1
- Application number: 14/729,423
- Filed: 2015-06-03
- Published: 2016-12-08
- Earliest claimed: 2015-06-03
- Assignees: Imediplus Inc.
- Applicants: Imediplus Inc.
- Inventors: Kun-Hsi TSAI; Yu TSAO; Shih-Hsuan KU; Tzu-Chen LIANG; Yun-Fan CHANG; Shih-I YANG
- Web page: https://patentgrep.com/patent/US20160354053A1

## CPC

- A61B 7/04: Electric stethoscopes
- A61B 5/02028: {Determining haemodynamic parameters not otherwise provided for, e.g. cardiac contractility or left ventricular ejection fraction}
- A61B 5/0803: {Recording apparatus specially adapted therefor}
- A61B 5/1107: {Measuring contraction of parts of the body, e.g. organ or muscle}
- A61B 5/4255: {Intestines, colon or appendix}
- A61B 5/7246: {using correlation, e.g. template matching or determination of similarity}
- A61B 5/7253: {characterised by using transforms}
- A61B 5/7264: {Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems}
- A61B 5/7267: {involving training the classification device}
- A61B 5/7275: {Determining trends in physiological measurement data; Predicting development of a medical condition based on physiological measurements, e.g. determining a risk factor}
- A61B 5/7282: {Event detection, e.g. detecting unique waveforms indicative of a medical condition (cough events A61B5/0823; seizures A61B5/4094; sleep apnoea A61B5/4818)}
- A61B 7/003: {Detecting lung or respiration noise}
- A61B 7/008: {Detecting noise of gastric tract, e.g. caused by voiding}
- A61B 8/0883: {for diagnosis of the heart}
- A61B 8/0891: {for diagnosis of blood vessels}
- A61B 8/5223: {for extracting a diagnostic or physiological parameter from medical diagnostic data (for algorithms to analyse biomedical images G06T7/0012)}
- G16H 40/63: for local operation
- G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems
- G16H 50/30: for calculating health indices; for individual health risk assessment
- A61B 5/0205: Simultaneously evaluating both cardiovascular conditions and different types of body conditions, e.g. heart and respiratory condition
- A61B 5/113: occurring during breathing
- A61B 2562/0204: Acoustic sensors
- A61N 1/36578: {controlled by mechanical motion of the heart wall, e.g. measured by an accelerometer or microphone}
- A61N 1/3904: {External heart defibrillators [EHD]}
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)
- G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- A61N: ELECTROTHERAPY; MAGNETOTHERAPY; RADIATION THERAPY; ULTRASOUND THERAPY (measurement of bioelectric currents A61B; surgical instruments, devices or methods for transferring non-mechanical forms of energy to or from the body A61B18/00; anaesthetic apparatus in general A61M; incandescent lamps H01K; infrared radiators for heating H05B)

## Abstract

A system for recognizing physiological sound comprises a receiving module, a feature extracting module, a classifier, and a comparing module. A method for recognizing physiological sound comprises receiving a physiological sound by the receiving module; extracting at least one feature from the physiological sound by the feature extraction module; classifying the at least one feature to identify at least one category by a classifier; and comparing the at least one category with a normal physiological sound and/or an abnormal physiological sound by the comparing module for evaluating a risk of disease. The method and system for recognizing physiological sound can precisely identify the specific physiological sound and exclude the noise.

## Claims

### Claim 1

A system for recognizing physiological sound, comprising:
a receiving module configured to receive a physiological sound;
a feature extracting module configured to extract at least one feature from the physiological sound; and
a classifier configured to classify the at least one feature to identify at least one category;
wherein the feature extracting module comprises a voice activity detector (VAD) module configured to detect at least one segment from the physiological sound, a Mel-frequency cepstrum coefficients (MFCC) module configured to transfer the at least one segment to at least one MFCC feature vector, and a K-means algorithm module configured to find at least one representative data point from the at least one MFCC feature vector.

### Claim 2 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the receiving module is a physiological recording device converting an analog signal of the physiological sound into a digital signal of the physiological sound.

### Claim 3 (depends on 2)

The system for recognizing physiological sound according to claim 2, wherein the physiological recording device is an electronic stethoscope.

### Claim 4

(canceled)

### Claim 5

(canceled)

### Claim 6

(canceled)

### Claim 7 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the classifier includes a K-nearest neighbor (KNN) module, a Gaussian mixture model (GMM) module, or a support vector machine (SVM) module.

### Claim 8 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the physiological sound includes heart sound, lung sound, bowel sound, vascular sound, tracheal breath sound, bronchioles breath sound or extremities sound.

### Claim 9 (depends on 8)

The system for recognizing physiological sound according to claim 8, wherein the heart sound comprises first heart sound (S1), second heart sound (S2) or a combination thereof.

### Claim 10 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the system further comprises:
a comparing module configured to compare the at least one category with a normal physiological sound and/or an abnormal physiological sound for evaluating a risk of disease.

### Claim 11 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the system further comprises an automated external defibrillator, a Holter monitor, a Cardiopulmonary Resuscitation (CPR) machine, a pacemaker, an implantable cardioverter defibrillator (ICD), an electrocardiogram (EKG), or an ultrasonic wave device.

### Claim 12 (depends on 1)

A method for recognizing physiological sound by the system as claimed in claim 1, comprising:
receiving a physiological sound by the receiving module;
extracting at least one feature from the physiological sound by the feature extraction module; and
classifying the at least one feature to identify at least one category by the classifier;
wherein the extracting at least one feature from the physiological sound by the feature extraction module comprises:
detecting at least one segment from the physiological sound by the VAD module;
transferring the at least one segment to at least one MFCC feature vector by the MFCC module;
finding at least one representative data point from the at least one MFCC feature vector by the K-means algorithm module.

### Claim 13

(canceled)

### Claim 14

(canceled)

### Claim 15

(canceled)

### Claim 16 (depends on 12)

The method for recognizing physiological sound according to claim 12, wherein the classifier includes a KNN module, a GMM module, or an SVM module.

### Claim 17 (depends on 12)

The method for recognizing physiological sound according to claim 12, wherein the physiological sound includes heart sound, lung sound, bowel sound, vascular sound, tracheal breath sound, bronchioles breath sound or extremities sound.

### Claim 18 (depends on 17)

The method for recognizing physiological sound according to claim 17, wherein the heart sound comprises first heart sound (S1), second heart sound (S2) or a combination thereof.

### Claim 19 (depends on 12)

The method for recognizing physiological sound according to claim 12, wherein the method further comprises:
comparing the at least one category with a normal physiological sound and/or an abnormal physiological sound by a comparing module for evaluating a risk of disease.

### Claim 20 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the classifier comprises a deep neural network (DNN) module.

### Claim 21 (depends on 20)

The system for recognizing physiological sound according to claim 20, wherein the DNN module comprises three hidden layers, and each hidden layer comprises 100 neurons.

### Claim 22 (depends on 9)

The system for recognizing physiological sound according to claim 9, wherein the feature extracting module is configured to calculating heart rates of the heart sound (S1) or the second heart sound (S2) and calculating a modulus of differences between two adjacent heart rates.

### Claim 23 (depends on 22)

The system for recognizing physiological sound according to claim 22, wherein the classifier is configured to identify regular heart sounds or irregular heart sound by comparing the modulus to 5% of a prior heart rate of the two adjacent heart rates, the regular heart sounds are identified when the modulus is smaller than 5% of a prior heart rate of the two adjacent heart rates, and the irregular heart sounds are identified when the modulus is higher than 5% of a prior heart rate of the two adjacent heart rates.

### Claim 24 (depends on 1)

The system for recognizing physiological sound according to claim 1, wherein the voice activity detector (VAD) module is configured for detecting the at least one segment from the physiological sound based on sound energy differences.

### Claim 25 (depends on 12)

The method for recognizing physiological sound according to claim 12, wherein the classifier comprises a deep neural network (DNN) module.

### Claim 26 (depends on 25)

The method for recognizing physiological sound according to claim 25, wherein the DNN module comprises three hidden layers, and each hidden layer comprises 100 neurons.

### Claim 27 (depends on 18)

The method for recognizing physiological sound according to claim 18, wherein the extracting at least one feature from the physiological sound by the feature extraction module comprises calculating heart rates of the heart sound (S1) or the second heart sound (S2) and calculating a modulus of differences between two adjacent heart rates.

### Claim 28 (depends on 27)

The method for recognizing physiological sound according to claim 27, wherein the heart rates are identified to be regular heart sounds or irregular heart sound by comparing the modulus to 5% of a prior heart rate of the two adjacent heart rates, the regular heart sounds are identified when the modulus is smaller than 5% of a prior heart rate of the two adjacent heart rates, and the irregular heart sounds are identified when the modulus is higher than 5% of a prior heart rate of the two adjacent heart rates.

### Claim 29 (depends on 12)

The method for recognizing physiological sound according to claim 12, wherein the voice activity detector (VAD) module is configured for detecting the at least one segment from the physiological sound based on sound energy differences.
