Cough sound automatic identification and analysis system and signal processing method

Through multi-dimensional feature parameter extraction and deep learning algorithms, the subjectivity and noise suppression problems of cough sound diagnosis in the existing technology are solved, comprehensive analysis of cough sound signals and accurate identification of different types of cough sounds, and diagnostic efficiency is improved .

CN119033360APending Publication Date: 2024-11-29CHENGDU MEDICAL COLLEGE
View PDF 0 Cites 1 Cited by

Patent Information

Application Number
CN202411288129.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-14
Publication Date
2024-11-29

AI Technical Summary

Technical Problem

The prior art has strong subjectivity and poor repetition in the diagnosis of cough sounds, which cannot effectively suppress noise, and lacks the ability to recognize different types of cough sounds, which limits its practicality in clinical diagnosis.

Method used

Multi-dimensional feature parameter extraction method, including time domain, frequency domain and intensity analysis, combined with deep learning algorithm, noise suppression and feature extraction are performed through Audacity and MATLAB, and deep learning algorithms are used to realize automatic recognition of cough sound patterns.

Benefits of technology

A comprehensive analysis of cough sound signals is achieved, and different types of cough sounds can be accurately collected and identified in noisy environments, improving diagnostic efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119033360A_ABST
    Figure CN119033360A_ABST
Patent Text Reader

Abstract

The invention provides a cough sound automatic identification and analysis system and a signal processing method. The cough sound automatic identification and analysis system comprises a cough sound acquisition device, a preprocessing module, a segmentation module and a feature extraction module, the cough sound collecting device is used for collecting cough sound signals at the position about 10 cm away from the face of the patient; the preprocessing module is used for carrying out noise suppression on the collected cough sound signals by utilizing Audacity software; the segmentation module is used for carrying out fragment segmentation on the preprocessed cough sound signals through amplitude analysis; and the feature extraction module is used for extracting time domain, frequency domain and intensity parameters of the segmented cough sound segments by utilizing MATLAB.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of automatic recognition and analysis of cough sounds, and in particular to an automatic recognition and analysis system for cough sounds and a signal processing method. Background Art

[0002] In the field of medical diagnosis, early identification and diagnosis of respiratory diseases are crucial. Coughing is one of the common symptoms of respiratory diseases, and its sound characteristics carry important information about the condition. However, traditional cough sound diagnosis methods mainly rely on doctors' auditory judgment, which has problems such as strong subjectivity and poor repeatability. With the development of digital signal processing and artificial intelligence technology, disease diagnosis by analyzing the characteristic parameters of cough sounds has become a new research direction.

[0003] The existing patent with publication number US20230368000A1 provides systems and methods for acoustic feature extraction and dual splitter models. The systems and methods of the present disclosure enable the detection and / or identification of signals in audio recordings using one or more signal segmentation techniques (including a computing system configured therefor). The computing system can receive a signal data signature of time-varying data, which has an event of interest, and segment the signal data signature into at least one segment of the time-varying data by utilizing a configured first hidden Markov model (HMM), by identifying a state change indicating the event of interest and a time change of at least one segment of the time-varying data and a segmented data of at least one segment of the time-varying event having a first length. The computing system can use a configured second HMM to segment at least one segment into sub-segments of the time-varying data by identifying a state change within at least one segment.

[0004] Currently, existing technologies usually only extract simple features of cough sounds, such as frequency and intensity, and ignore acoustic features in more dimensions, resulting in an incomplete and in-depth analysis of cough sounds. In a noisy environment, it is difficult to effectively suppress noise, which affects the accurate collection and feature analysis of cough sounds. In addition, the algorithms relied on by existing cough recognition systems cannot accurately distinguish different types of cough sounds, limiting their practicality in clinical diagnosis. Many solutions still rely on manual analysis and lack intelligent and automated processing procedures, resulting in low diagnostic efficiency. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides an automatic cough recognition and analysis system and a signal processing method, which realizes a comprehensive analysis of cough sound signals by extracting multi-dimensional cough sound feature parameters including time domain, frequency domain and intensity.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions: A cough sound analysis system, comprising: A cough sound collection device is used to collect cough sound signals at a distance of about 10 cm from the patient's face; The pre-processing module uses Audacity software to suppress the noise of the collected coughing sound signal; Segmentation module, which segments the preprocessed cough signal through amplitude analysis; The feature extraction module uses MATLAB to extract the time domain, frequency domain and intensity parameters of the segmented cough sound segments.

[0007] Preferably, the preprocessing module further comprises a Fourier transform unit for performing frequency domain conversion on the cough signal, and a filtering unit for removing background noise by filtering technology.

[0008] Preferably, the feature extraction module includes: A time domain analysis unit, used to calculate the energy, root mean square value and duration of the cough sound signal; A frequency domain analysis unit, used for calculating the spectrum energy, spectrum skewness and spectrum amplitude of the cough signal; The intensity analysis unit is used to calculate the sound pressure level of the cough signal.

[0009] A signal processing method for analyzing cough sounds comprises the following steps: Place the recording device about 10 cm above the patient's nose to collect coughing signals; Preprocess the collected cough signal to remove noise; The cough signal is segmented into multiple segments through amplitude analysis; The time domain, frequency domain and intensity parameters of the segmented cough sound segments are extracted.

[0010] Preferably, the pretreatment step comprises: Perform short-time Fourier transform on the cough signal to obtain frequency domain information; Filtering techniques are used to remove irrelevant noise signals.

[0011] Preferably, the characteristic parameters include energy, root mean square value, duration, spectrum energy, spectrum skewness, spectrum amplitude and sound pressure level.

[0012] A cough sound automatic recognition system, comprising: A collection module, used for collecting coughing signals of patients; A processing module, connected to the acquisition module, for preprocessing the acquired cough signal and extracting characteristic parameters; The recognition module uses deep learning algorithms to analyze and classify the extracted feature parameters to achieve automatic recognition of cough sound patterns.

[0013] Preferably, the recognition module improves the accuracy and robustness of cough sound classification by integrating multiple neural network models.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. The present invention aims to extract multi-dimensional cough feature parameters including time domain, frequency domain and intensity, so as to achieve a comprehensive analysis of cough signal.

[0015] 2. By adopting advanced signal processing technology, the present invention can accurately collect cough sounds in various environments and effectively suppress noise even in noisy backgrounds.

[0016] 3. The present invention uses intelligent algorithms such as deep learning to improve the ability to recognize cough patterns and can more accurately distinguish different types of coughs.

[0017] 4. By integrating artificial intelligence technology, the present invention realizes fully automated processing of cough sound feature analysis and improves diagnostic efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0019] Figure 1 It is a schematic diagram of the framework of the cough sound analysis system of the present invention; Figure 2 This is a schematic diagram of the cough sound collection process of the present invention; Figure 3 Schematic diagram of the cough sound preprocessing process of the present invention; Figure 4 This is a schematic diagram of the cough sound signal before noise reduction of the present invention; Figure 5 This is a schematic diagram of a cough sound signal after noise reduction according to the present invention; Figure 6 It is a schematic diagram of cough sound segmentation signal of the present invention. DETAILED DESCRIPTION

[0020] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art can think of other obvious variations. The basic principles of the present invention defined in the following description can be used for other embodiments, variations, improvements, equivalents, and other technical solutions that do not deviate from the spirit and scope of the present invention.

[0021] Those skilled in the art should understand that, in the disclosure of the present invention, the directions or positions indicated by the terms "longitudinal", "lateral", "up", "down", "left", "right", "front", "back", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc. are based on the directions or positional relationships shown in the accompanying drawings, which are only simplified descriptions for the convenience of describing the present invention, and do not indicate or imply that the device or component referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, the above terms should not be understood as limitations on the present invention.

[0022] It is to be understood that the term "one" should be understood as "at least one" or "one or more", that is, in one embodiment, the number of an element may be one, while in another embodiment, the number of the element may be multiple, and the term "one" should not be understood as a limitation on the quantity.

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings: See also Figure 1 ,The present invention provides a cough sound analysis system, including a cough sound collection device, a preprocessing module, a segmentation module, and a feature extraction module; The cough sound collection device is used to collect cough sound signals at a distance of about 10 cm from the patient's face; the preprocessing module uses Audacity software to suppress noise on the collected cough sound signals; the segmentation module segments the preprocessed cough sound signals through amplitude analysis; the feature extraction module uses MATLAB to extract the time domain, frequency domain and intensity parameters of the segmented cough sound segments.

[0024] See also Figure 2 The cough sound collection device of the present invention collects cough sound signals through the following steps: 1) Place the recording device (mobile phone) horizontally above the patient's nose - at the root of the nose (the trajectory of saliva and droplets is a downward parabola, so as to avoid contamination of the recorder's hands when recording the coughing sound); 2) Make sure the phone is placed about 1 cm away from the patient's face to ensure good sound recording. The data collector should tell the subject the gesture to start coughing so that the patient knows when to start coughing and ensure smooth recording of the coughing sound; 3) When recording coughing sounds, the collector should place the phone in the standard position, press the record button to start recording, and then signal the patient to start coughing - to ensure the integrity of the coughing sound recording; 4) The length of the recording time depends on the number of times the patient coughs. The default cough sound is 2s. Each patient is recorded for 4 consecutive days, three times a day (8:00-10:00, 10:00-12:00, 14:30-17:00). Each person should have 12 audios after the recording is completed.

[0025] See also Figure 3-Figure 5 , the preprocessing module of the present invention is used to increase the lung sound signal and eliminate the noise signal; Audacity software is further used to eliminate the interference of noise signals, which uses time and frequency transformation to separate these signals. Time transformation usually includes segmentation and smoothing to remove noise and unnecessary signals; frequency transformation usually includes converting signals to the frequency domain using methods such as Fourier transform, wavelet transform, Hilbert-Huang transform, etc.; and then filtering or extracting useful signals by removing non-correlated signals.

[0026] Further sound enhancement uses filters to remove unnecessary noise, or uses algorithms based on waveform amplitude or spectral characteristics to enhance useful lung sound signals.

[0027] See also Figure 6 The segmentation module of the present invention separates the cough sound segments according to the amplitude after the cough sound is preprocessed.

[0028] The feature extraction module of the present invention comprises: A time domain analysis unit, used to calculate the energy, root mean square value and duration of the cough sound signal; A frequency domain analysis unit, used for calculating the spectrum energy, spectrum skewness and spectrum amplitude of the cough signal; The intensity analysis unit is used to calculate the sound pressure level of the cough signal.

[0029] The feature extraction module uses MATLAB to extract feature parameters of the standard cough sound fragments from the time domain, frequency domain and intensity. The feature extraction methods include short-time Fourier transform, wavelet packet transform, wavelet transform, cepstral coefficients and other methods to extract feature parameters with time domain, frequency domain and intensity as the main parameters.

[0030] (1) The time domain analysis unit analyzes and calculates the waveform in the audio time domain to obtain the relevant time domain characteristic parameters, including: Energy, RMS (root mean square value), Duration, etc., where Energy represents the overall strength of the cough signal in the time domain, which is calculated by squaring and summing the amplitude of the signal; RMS refers to the root mean square value of the signal over a period of time, representing the overall energy or amplitude of the signal, that is, describing the stability of a section of the signal; Duration, as a time domain characteristic parameter of audio, provides information about the signal in time, which is of great significance for audio processing, analysis and recognition. The duration can be calculated and obtained through a signal with clear start and end marks.

[0031] (2) The frequency domain analysis unit performs a short-time Fourier transform on the time domain signal to obtain the signal's spectrum information, including spectrum energy, spectrum skewness, spectrum amplitude, etc., that is, the components of the signal at different frequencies. These frequency domain characteristic parameters provide information about the structure and characteristics of the cough signal in the frequency domain. They are used together with the time domain and intensity characteristic parameters to fully understand all aspects of the cough sound.

[0032] (3) The intensity analysis unit performs calculations based on the time domain or frequency domain combined with Fourier transform or other related technologies, including the measurement of sound pressure level. Sound pressure level is a physical quantity that describes the intensity of sound. It is related to the amplitude and energy of the sound and is usually expressed in decibels (dB).

[0033] The signal processing method based on the above cough sound analysis system comprises the following steps: Place the recording device about 10 cm above the patient's nose to collect coughing signals; Preprocess the collected cough signal to remove noise; The cough signal is segmented into multiple segments through amplitude analysis; The time domain, frequency domain and intensity parameters of the segmented cough sound segments are extracted.

[0034] The preprocessing steps include: Perform short-time Fourier transform on the cough signal to obtain frequency domain information; Filtering techniques are used to remove irrelevant noise signals.

[0035] The automatic cough recognition system based on the cough analysis system includes a collection module, a processing module and a recognition module; A collection module, used for collecting coughing signals of patients; A processing module, connected to the acquisition module, for preprocessing the acquired cough signal and extracting characteristic parameters; The recognition module uses deep learning algorithms to analyze and classify the extracted feature parameters to achieve automatic recognition of cough sound patterns.

[0036] Among them, the recognition module improves the accuracy and robustness of cough sound classification by integrating multiple neural network models.

[0037] Those skilled in the art should understand that the embodiments of the present invention described above and shown in the accompanying drawings are only examples and do not limit the present invention. The purpose of the present invention has been fully and effectively achieved. The functions and structural principles of the present invention have been demonstrated and explained in the embodiments, and the embodiments of the present invention may be deformed or modified in any way without departing from the principles.

Claims

1. A cough sound analysis system, characterized in that: include: A cough sound collection device is used to collect cough sound signals at a distance of about 10 cm from the patient's face; The pre-processing module uses Audacity software to suppress the noise of the collected coughing sound signal; Segmentation module, which segments the preprocessed cough signal through amplitude analysis; The feature extraction module uses MATLAB to extract the time domain, frequency domain and intensity parameters of the segmented cough sound segments.

2. A cough sound analysis system according to claim 1, characterized in that: The preprocessing module further includes a Fourier transform unit for performing frequency domain conversion on the cough signal, and a filtering unit for removing background noise by filtering technology.

3. A cough sound analysis system according to claim 2, characterized in that: The feature extraction module comprises: A time domain analysis unit, used to calculate the energy, root mean square value and duration of the cough sound signal; A frequency domain analysis unit, used for calculating the spectrum energy, spectrum skewness and spectrum amplitude of the cough signal; The intensity analysis unit is used to calculate the sound pressure level of the cough signal.

4. A signal processing method for analyzing cough sound, characterized in that: The following steps are involved: Place the recording device about 10 cm above the patient's nose to collect coughing signals; Preprocess the collected cough signal to remove noise; The cough signal is segmented into multiple segments through amplitude analysis; The time domain, frequency domain and intensity parameters of the segmented cough sound segments are extracted.

5. The signal processing method according to claim 4, characterized in that: The pre-processing step comprises: Perform short-time Fourier transform on the cough signal to obtain frequency domain information; Filtering techniques are used to remove irrelevant noise signals.

6. The signal processing method according to claim 5, characterized in that: The characteristic parameters include energy, root mean square value, duration, spectrum energy, spectrum skewness, spectrum amplitude and sound pressure level.

7. A cough sound automatic recognition system, characterized by: include, A collection module, used for collecting coughing signals of patients; A processing module, connected to the acquisition module, for preprocessing the acquired cough signal and extracting characteristic parameters; The recognition module uses deep learning algorithms to analyze and classify the extracted feature parameters to achieve automatic recognition of cough sound patterns.

8. The automatic cough recognition system according to claim 7, characterized in that: The recognition module improves the accuracy and robustness of cough sound classification by integrating multiple neural network models.

Citation Information

Cited By

  • Live pig cough audio identification method and system based on improved U-Net and PANNS-CNN6

    CN119673179A