CN111248938A - Real-time heart sound signal diagnosis system and diagnosis method thereof - Google Patents
Real-time heart sound signal diagnosis system and diagnosis method thereof Download PDFInfo
- Publication number
- CN111248938A CN111248938A CN202010110855.5A CN202010110855A CN111248938A CN 111248938 A CN111248938 A CN 111248938A CN 202010110855 A CN202010110855 A CN 202010110855A CN 111248938 A CN111248938 A CN 111248938A
- Authority
- CN
- China
- Prior art keywords
- heart sound
- diagnosis
- real
- time
- heart
- 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.)
- Pending
Links
Images
Classifications
-
- A—HUMAN NECESSITIES
- A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
- A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
- A61B7/00—Instruments for auscultation
- A61B7/02—Stethoscopes
- A61B7/04—Electric stethoscopes
Landscapes
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Engineering & Computer Science (AREA)
- Biomedical Technology (AREA)
- Heart & Thoracic Surgery (AREA)
- Medical Informatics (AREA)
- Molecular Biology (AREA)
- Surgery (AREA)
- Animal Behavior & Ethology (AREA)
- General Health & Medical Sciences (AREA)
- Public Health (AREA)
- Veterinary Medicine (AREA)
- Measuring Pulse, Heart Rate, Blood Pressure Or Blood Flow (AREA)
Abstract
The invention discloses a real-time heart sound diagnosis system and a diagnosis method thereof. The real-time heart sound diagnosis method comprises the following steps: receiving real-time heart sound data and performing framing processing; reading the heart sound data frame by frame, carrying out noise reduction and filtering processing, and then caching; extracting the heart sound envelopes of the cached noise-reduced and filtered heart sound data through a first algorithm, and screening out the heart sound envelopes which accord with physiological characteristics; extracting a characteristic parameter set of the heart sound envelope which accords with the physiological characteristics; and inputting the characteristic parameter set of the heart sound envelope into a diagnosis model by taking the characteristic parameter set of the heart sound envelope as an input node to obtain a diagnosis result. The invention can obtain accurate and real-time heart sound diagnosis results and provide more convenient conditions for monitoring the physical health of people.
Description
Technical Field
The invention relates to a heart sound signal diagnosis technology, in particular to a real-time heart sound signal diagnosis system and a diagnosis method thereof.
Background
With the increase of population, aging is serious, medical resources are gradually strained, and the heart is one of the most important organs of human body, and the health problem of the heart is particularly attractive. At present, the initial diagnosis of the heart problem is usually confirmed by manual auscultation, and the manual auscultation mode is influenced by various factors such as the surrounding environment, the quality of a stethoscope, the experience of a doctor and the like, so that the uncertainty is high.
In the prior art, a training database for heart sound diagnosis is established in a cloud server, and the heart sound data uploaded by a user is analyzed by a diagnosis algorithm, so that whether the heart sound is abnormal or not is judged. However, this approach has major disadvantages. For example: the existing scheme depends on a user to transmit data to a cloud server through the Internet or a mobile network, the result is fed back to the user after the cloud carries out diagnosis and analysis, the diagnosis process needs to depend on the network, the network transmission efficiency and the diagnosis time consumption cause that the real-time performance is not good enough, and the convenience is not enough. In addition, if the collected heart sound data has large environmental noise, the accuracy and stability of the diagnosis result are difficult to ensure. Moreover, the algorithm on the server is not perfect enough, only the uploaded sound is analyzed, and many features except the heart sound are mixed into the feature parameter set obtained by the algorithm.
Disclosure of Invention
The invention provides a real-time heart sound diagnosis system and a diagnosis method thereof, aiming at solving the technical problem of inaccurate network diagnosis in the prior art.
The invention provides a real-time heart sound diagnosis method, which comprises the following steps:
step 1, receiving real-time heart sound data and performing framing processing;
step 2, reading the heart sound data frame by frame, performing noise reduction and filtering processing, and caching;
step 3, extracting the heart sound envelopes of the heart sound data subjected to noise reduction and filtering processing through a first algorithm, and screening out the heart sound envelopes which accord with physiological characteristics;
step 4, extracting a characteristic parameter set of the heart sound envelope conforming to the physiological characteristics;
and 5, inputting the characteristic parameter set of the heart sound envelope into a diagnosis model by taking the characteristic parameter set of the heart sound envelope as an input node to obtain a diagnosis result.
Further, the first algorithm adopts one of a shannon envelope algorithm, a hilbert-yellowing transform algorithm or a wavelet transform algorithm.
Further, the heart sound envelopes which accord with the physiological characteristics are screened out by judging whether the time length of one heart sound envelope accords with a preset time length range.
Further, after the received real-time heart sound data is subjected to framing processing, the noise reduction algorithm provided by the webpage instant messaging technology is adopted to perform frame-by-frame noise reduction processing on the heart sound data.
Further, each frame of the heart sound data after framing comprises at least one heart sound.
Further, the diagnosis model is a BPNN neural network diagnosis model, and the BPNN neural network diagnosis model is constructed through BPNN neural network parameters obtained from a remote server.
Further, the BPNN neural network parameters are obtained by the following steps:
obtaining a characteristic parameter set of sample data by adopting the steps 1 to 4;
attaching corresponding pathological labels to the characteristic parameter set of the sample data;
and taking the characteristic parameter set as an input node and the corresponding pathological label as an output node, and performing iterative training with pathology as classification to obtain the BPNN neural network parameters.
The real-time heart sound diagnosis system provided by the invention adopts the diagnosis method of real-time heart sound in the technical scheme for diagnosis, and the real-time heart sound diagnosis system specifically comprises the following steps:
the electronic stethoscope is used for acquiring the heart sound data;
the diagnosis module is arranged on the intelligent terminal and used for receiving and processing the heart sound data sent by the electronic stethoscope and obtaining a diagnosis result according to the diagnosis model;
and the remote server is used for providing the construction parameters of the diagnosis model.
Further, the construction parameters include the BPNN neural network parameters.
Further, the electronic stethoscope is communicated with the diagnosis module through a low-power Bluetooth technology.
The invention can solve the problem of accuracy of off-line intelligent heart sound diagnosis, well inhibits various noises generated in the recording of the stethoscope, greatly improves the effectiveness and pertinence of the heart sound characteristic parameters, and further improves the accuracy of the neural network technology in the aspect of heart sound diagnosis. The stethoscope collects the pathological signals of heart sounds, and the APP of the diagnosis module is used for carrying out real-time noise reduction on the signals, filtering partial noise, further carrying out high-pass filtering on the signals, and filtering high-energy low-frequency noise below 60Hz, so that the frequency band of the heart sounds is highlighted. And then, performing cloud collection on the processed heart sound data serving as sample data, performing effectiveness judgment on the heart sound envelope by using a Shannon envelope algorithm, finding out a heart sound segment of a complete period, so that the heart sound segment does not contain fricatives or environmental noise with high energy, setting the frequency band range of the MFCC filtering group to enable the frequency band (60-600 Hz) of the MFCC filtering group to be in accordance with the heart sound, performing characteristic coefficient extraction, entering a BPNN neural network for deep learning training, and outputting corresponding construction parameters. The intelligent terminal can download the construction parameters from the server, in the auscultation process, the app of the diagnosis module receives sound collection of the electronic stethoscope in real time, noise reduction and filtering processing is carried out to obtain heart sound fragments with fricatives and strong environmental noises eliminated, the frequency band range of the MFCC filtering group is set to enable the frequency band to be in accordance with the heart sound (60-600 Hz) for characteristic coefficient extraction, the BPNN neural network is reconstructed by combining preset or neural network parameters issued by the server to carry out classification prediction, and the offline intelligent diagnosis function is achieved. Therefore, the method is not easily limited by a network during diagnosis, and better diagnosis recognition rate can be obtained by periodically upgrading, updating and optimizing the neural network model parameters through the network, so that the product can be conveniently iterated, and the user experience is improved.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without inventive exercise.
FIG. 1 is an overall architecture diagram of the present invention.
FIG. 2 is a flow chart of the diagnostic and model training process of the present invention.
Detailed Description
In order to make the technical problems, technical solutions and advantageous effects to be solved by the present invention more clearly apparent, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The BP neural network-based machine learning algorithm constructs heart sound characteristics through an effective preprocessing method, performs learning training, continuously iterates to correct parameters, establishes a set of offline diagnosis mechanism giving more than 90% accuracy to heart sounds, greatly improves the convenience and accuracy of equipment use, and enables the electronic stethoscope to be widely applied to scenes such as medical treatment, teaching, old age, household and the like. The principles and construction of the present invention will be described in detail below with reference to the drawings and examples.
Fig. 1 shows a block diagram of the architecture of the present invention. The real-time heart sound diagnosis system comprises an electronic stethoscope, a diagnosis module and a remote server.
The electronic stethoscope is a heart sound collecting device and is used for collecting heart beating sound signals of a user within a period of time, converting the heart beating sound signals into digital signal data, and transmitting the real-time heart sound data to the intelligent terminal through a Low-power Bluetooth (BLE) technology to complete data exchange and communication between hardware and software.
The diagnosis module is installed on intelligent terminal, installs promptly on the mobile terminal, and intelligent terminal includes but not limited to smart mobile phone, intelligent wrist-watch, panel computer, notebook computer and desktop etc. and the diagnosis module can utilize intelligent terminal's bluetooth module to receive the real-time heart sound data of electron stethoscope. The specific expression form of the diagnosis module can be an application program (APP), the APP runs in an intelligent terminal, and the APP is mainly used for heart sound data processing, including filtering, noise reduction, Shannon envelope detection, fricative sound, environmental sound and other exclusion detection, MFCC feature extraction of heart sound, BPNN neural network reconstruction, prediction, diagnosis result feedback and the like. The diagnostic module based on the diagnostic result feedback diagnostic model may use the construction parameters stored in the memory of the intelligent terminal, and the diagnostic module constructs the diagnostic model for diagnosis through the construction parameters.
The remote server is mainly used for storing heart sound data and training a diagnosis model, the BPNN neural network model is continuously trained and updated along with the expansion of sample data collected by the remote server, so that the diagnosis accuracy can be further improved, the obtained data (such as BPNN neural network parameters) can be shared with mechanisms such as hospitals and the like, so that a more professional and convenient remote inquiry and consultation service is provided for users, the remote server and the intelligent terminal can be connected through a network, so that the intelligent terminal can download corresponding latest BPNN neural network parameters from the remote server, and the diagnosis model can be smoothly constructed. The remote server may also provide the diagnostic module with the latest parameter configuration.
Fig. 2 is a method of diagnosing real-time heart sounds according to the present invention.
When the buffered heart sound data is sufficiently framed, the diagnostic module performs framing processing so that each frame of heart sound data contains at least one heart sound (the sound of the heart beating, usually the heart sounds are generated twice by the heart beating once S1 and S2), namely at least one heart sound S1 and/or at least one heart sound S2. For example, the frame length may be set to 10ms, facilitating subsequent processing.
And reading the framed heart sound data frame by frame, performing noise reduction and filtering processing, and then caching. Because the heart sounds collected by the stethoscope generally include noise introduced by various circuits, components and the like and weak noise in the environment, the diagnostic module performs frame-by-frame processing by using a real-time noise reduction algorithm, preferably, a noise reduction algorithm (also called webrtc noise reduction algorithm) provided by a webpage instant messaging technology can be used for processing, and noise mixed in the heart sounds can be well suppressed. Considering that the frequency range of the heart sound acquired by the stethoscope is 60-600 Hz, high-pass filtering processing needs to be carried out on high-energy low-frequency components generated in a circuit, sound components below 60Hz can be effectively inhibited by setting reasonable roll-off coefficients and high-pass frequency points, so that the heart sound is highlighted, a biquad filtering algorithm can be typically adopted for frame-by-frame processing and caching, and the heart sound can be specifically cached in an FIFO memory of an intelligent terminal.
Accumulating to a certain time length (such as 1s, etc.), then carrying out envelope transformation on the cached heart sound data, and extracting the heart sound envelope of the cached noise-reduced and filtered heart sound data through a first algorithm, wherein the specific first algorithm can be any one of a shannon envelope algorithm, a hilbert-yellowing transform algorithm or a wavelet transform algorithm. And then screening out a heart sound envelope conforming to physiological characteristics, specifically, identifying the length of the heart sound by identifying the heart sound envelope, wherein the typical length of the first heart sound S1 or the second heart sound S2 does not exceed 200ms, frictional noise or strong environmental noise generated by a stethoscope head of a general stethoscope exceeds the length, checking the effectiveness of the heart sound envelope by using the length, judging the frictional noise or the environmental noise if the length exceeds 200ms, discarding segments, returning to the initial step, continuously collecting real-time heart sound data, and performing subsequent processing and use until the heart sound envelope in the obtained heart sound data conforms to the physiological characteristics.
The method comprises the steps of extracting a feature parameter set of a heart sound envelope according with physiological features, specifically carrying out MFCC algorithm processing on the heart sound envelope, setting a frequency band range of an MFCC filtering group, typically setting the frequency band range between 60Hz and 600Hz, enabling the feature parameter set to accord with the physiological features of heart sounds, and constructing an accurate feature parameter set. And then reconstructing a BPNN neural network diagnosis model on the intelligent terminal by using BPNN neural network parameters preset in the intelligent terminal or updated from a remote server, and taking the extracted MFCC characteristic parameter set as an input node of the BPNN neural network diagnosis model to obtain a diagnosis result, thereby realizing pathological diagnosis of the heart sound.
The above is the diagnostic process, and the specific construction parameters need the remote server to collect enough sample data in advance to train to obtain. Taking BPNN neural network parameters as an example, a remote server carries out frame processing on sample data in advance by adopting the same method, reads corresponding heart sound data frame by frame, carries out noise reduction and filtering processing and then caches the data, extracts the heart sound envelopes of the cached noise-reduced and filtered heart sound data through a first algorithm, screens out the heart sound envelopes conforming to physiological characteristics, extracts characteristic parameter sets conforming to the physiological characteristics of the heart sound envelopes, and attaches corresponding pathological labels to the characteristic parameter sets. And taking the characteristic parameter sets as input nodes and corresponding pathological labels as output nodes, and performing iterative training with pathology as classification to finally obtain the BPNN neural network parameters. With the continuous accumulation of sample data, the remote server can regularly update the parameters of the BPNN neural network, so that the result of the diagnosis model is more and more accurate.
The invention realizes the communication between the electronic stethoscope and the mobile terminal (intelligent terminal) by using the BLE technology, and has the advantages of fast connection and low energy consumption. Except for Bluetooth, the electronic stethoscope can be communicated with the mobile terminal in a wifi mode or the like. The invention sets the number of the filter sets to be 60-600 according to the heart sound frequency characteristics, so that the signal characteristics extracted by the MFCC can better accord with the judgment experience of professionals. The invention can improve the accuracy of the diagnosis result by real-time denoising, real-time filtering and real-time effectiveness detection processing by using shannon envelope, and it is to be noted that the algorithm adopted by the invention is not limited to the shannon algorithm, and also can use hilbert-yellow transform, wavelet transform and the like, and the shannon envelope algorithm is only one of the optimal algorithms adopted by the invention. Because the diagnosis model can be built in time through the building parameters, the BPNN neural network can be used for diagnosing and processing the heart sound data locally under the condition of no networking, and the diagnosis result can be provided conveniently and quickly. Furthermore, through the updating iteration of the sample data training of the remote server, the mobile terminal can update the corresponding parameter configuration of the diagnosis module through the network, and the flexibility is improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (10)
1. A method for real-time diagnosis of heart sounds, comprising:
step 1, receiving real-time heart sound data and performing framing processing;
step 2, reading the heart sound data frame by frame, performing noise reduction and filtering processing, and caching;
step 3, extracting the heart sound envelopes of the heart sound data subjected to noise reduction and filtering processing through a first algorithm, and screening out the heart sound envelopes which accord with physiological characteristics;
step 4, extracting a characteristic parameter set of the heart sound envelope conforming to the physiological characteristics;
and 5, inputting the characteristic parameter set of the heart sound envelope into a diagnosis model by taking the characteristic parameter set of the heart sound envelope as an input node to obtain a diagnosis result.
2. The method of diagnosing heart sounds in real time as set forth in claim 1, wherein the first algorithm employs one of a shannon envelope algorithm, a hilbert-yellow transform algorithm, or a wavelet transform algorithm.
3. The method for real-time diagnosis of heart sounds according to claim 1, wherein the heart sound envelopes corresponding to the physiological characteristics are selected by determining whether a time length of one of the heart sound envelopes corresponds to a predetermined time length range.
4. The method for diagnosing real-time heart sounds according to claim 1, wherein the received real-time heart sound data is de-noised frame by frame using a de-noising algorithm provided by a web instant messaging technique after being processed in frames.
5. The method of claim 1, wherein the segmented heart sound data comprises at least one heart sound for each frame.
6. The method according to any one of claims 1 to 5, wherein the diagnostic model is a BPNN neural network diagnostic model, and the BPNN neural network diagnostic model is constructed by using BPNN neural network parameters acquired from a remote server.
7. The method of real-time heart sound diagnosis of claim 6, wherein the BPNN neural network parameters are obtained by:
obtaining a characteristic parameter set of sample data by adopting the steps 1 to 4;
attaching corresponding pathological labels to the characteristic parameter set of the sample data;
and taking the characteristic parameter set as an input node and the corresponding pathological label as an output node, and performing iterative training with pathology as classification to obtain the BPNN neural network parameters.
8. A real-time heart sound diagnosis system which performs diagnosis using the method for diagnosing real-time heart sounds according to any one of claims 1 to 7, comprising:
the electronic stethoscope is used for acquiring the heart sound data;
the diagnosis module is arranged on the intelligent terminal and used for receiving and processing the heart sound data sent by the electronic stethoscope and obtaining a diagnosis result according to the diagnosis model;
and the remote server is used for providing the construction parameters of the diagnosis model.
9. The real-time heart sound diagnostic system of claim 8, wherein the build parameters include the BPNN neural network parameters.
10. The real-time heart sound diagnostic system of claim 8, wherein the electronic stethoscope communicates with the diagnostic module via low-power bluetooth technology.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010110855.5A CN111248938A (en) | 2020-02-24 | 2020-02-24 | Real-time heart sound signal diagnosis system and diagnosis method thereof |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010110855.5A CN111248938A (en) | 2020-02-24 | 2020-02-24 | Real-time heart sound signal diagnosis system and diagnosis method thereof |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111248938A true CN111248938A (en) | 2020-06-09 |
Family
ID=70943647
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010110855.5A Pending CN111248938A (en) | 2020-02-24 | 2020-02-24 | Real-time heart sound signal diagnosis system and diagnosis method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111248938A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112971802A (en) * | 2021-02-08 | 2021-06-18 | 中北大学 | Heart sound signal detection method and system based on deep learning model |
| CN113768532A (en) * | 2021-08-20 | 2021-12-10 | 中北大学 | Health detection method and system based on five-path heart sound signal classification algorithm |
| CN113974679A (en) * | 2020-11-13 | 2022-01-28 | 广东科学技术职业学院 | Stethoscope, heart sound detection method, heart sound detector and auscultation system |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103479383A (en) * | 2013-09-25 | 2014-01-01 | 清华大学 | Method and device for analyzing heart sound signals, and intelligent heart stethoscope provided with device for analyzing heart sound signals |
| CN105249986A (en) * | 2015-09-30 | 2016-01-20 | 哈尔滨师范大学 | Heart sound signal period parameter estimating method |
| CN106022258A (en) * | 2016-05-18 | 2016-10-12 | 成都济森科技有限公司 | Digital stethoscope and method for filtering heart sounds and extracting lung sounds |
| CN107945817A (en) * | 2017-11-15 | 2018-04-20 | 广东顺德西安交通大学研究院 | Heart and lung sounds signal sorting technique, detection method, device, medium and computer equipment |
| CN109965900A (en) * | 2019-02-02 | 2019-07-05 | 大连交通大学 | A kind of intelligent mothers and sons suit suitable for baby-monitoring and see maintaining method |
| CN110353725A (en) * | 2019-07-10 | 2019-10-22 | 东南大学 | A kind of heart sound acquisition and analysis system and method based on cloud framework |
| CN110720946A (en) * | 2019-09-26 | 2020-01-24 | 南通大学 | Intelligent auscultation system based on deep learning |
| CN110755108A (en) * | 2019-11-04 | 2020-02-07 | 合肥望闻健康科技有限公司 | Heart sound classification method, system and device based on intelligent stethoscope and readable storage medium |
-
2020
- 2020-02-24 CN CN202010110855.5A patent/CN111248938A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103479383A (en) * | 2013-09-25 | 2014-01-01 | 清华大学 | Method and device for analyzing heart sound signals, and intelligent heart stethoscope provided with device for analyzing heart sound signals |
| CN105249986A (en) * | 2015-09-30 | 2016-01-20 | 哈尔滨师范大学 | Heart sound signal period parameter estimating method |
| CN106022258A (en) * | 2016-05-18 | 2016-10-12 | 成都济森科技有限公司 | Digital stethoscope and method for filtering heart sounds and extracting lung sounds |
| CN107945817A (en) * | 2017-11-15 | 2018-04-20 | 广东顺德西安交通大学研究院 | Heart and lung sounds signal sorting technique, detection method, device, medium and computer equipment |
| CN109965900A (en) * | 2019-02-02 | 2019-07-05 | 大连交通大学 | A kind of intelligent mothers and sons suit suitable for baby-monitoring and see maintaining method |
| CN110353725A (en) * | 2019-07-10 | 2019-10-22 | 东南大学 | A kind of heart sound acquisition and analysis system and method based on cloud framework |
| CN110720946A (en) * | 2019-09-26 | 2020-01-24 | 南通大学 | Intelligent auscultation system based on deep learning |
| CN110755108A (en) * | 2019-11-04 | 2020-02-07 | 合肥望闻健康科技有限公司 | Heart sound classification method, system and device based on intelligent stethoscope and readable storage medium |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113974679A (en) * | 2020-11-13 | 2022-01-28 | 广东科学技术职业学院 | Stethoscope, heart sound detection method, heart sound detector and auscultation system |
| CN112971802A (en) * | 2021-02-08 | 2021-06-18 | 中北大学 | Heart sound signal detection method and system based on deep learning model |
| CN113768532A (en) * | 2021-08-20 | 2021-12-10 | 中北大学 | Health detection method and system based on five-path heart sound signal classification algorithm |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111248938A (en) | Real-time heart sound signal diagnosis system and diagnosis method thereof | |
| CN112164408A (en) | Pig coughing sound monitoring and early warning system based on deep learning | |
| CN109044280B (en) | Sleep staging method and related equipment | |
| Accou et al. | Modeling the relationship between acoustic stimulus and EEG with a dilated convolutional neural network | |
| CN111685774B (en) | OSAHS Diagnosis Method Based on Probability Integrated Regression Model | |
| Roy et al. | MIEEG-GAN: generating artificial motor imagery electroencephalography signals | |
| CN112869716B (en) | Pulse feature identification system and method based on two-channel convolutional neural network | |
| CN202801659U (en) | Electronic stethoscope based on intelligent identification function | |
| Sujadevi et al. | A hybrid method for fundamental heart sound segmentation using group-sparsity denoising and variational mode decomposition | |
| CN113116300A (en) | Physiological signal classification method based on model fusion | |
| CN114190944A (en) | Robust emotion recognition method based on electroencephalogram signals | |
| CN113436726A (en) | Automatic lung pathological sound analysis method based on multi-task classification | |
| CN111789629A (en) | Breath sound intelligent diagnosis and treatment system and method based on deep learning | |
| CN110415824B (en) | Cerebral apoplexy disease risk assessment device and equipment | |
| Lee et al. | Restoration of lung sound signals using a hybrid wavelet-based approach | |
| Jain et al. | An adaptive method for shrinking of wavelet coefficients for phonocardiogram denoising | |
| Talbi | New approach of ECG denoising based on 1-D double-density complex DWT and SBWT | |
| Jain et al. | A robust algorithm for segmentation of phonocardiography signal using tunable quality wavelet transform | |
| CN114488841B (en) | Data collection processing method of intelligent wearable device | |
| Morshed et al. | Automated heart valve disorder detection based on PDF modeling of formant variation pattern in PCG signal | |
| CN113033358A (en) | Cuff-free blood pressure modeling method based on deep neural network | |
| Malhotra et al. | Electrocardiogram signals denoising using improved variational mode decomposition | |
| CN111150422A (en) | Intelligent auscultation system and auscultation method thereof | |
| Liutkus et al. | Source separation for target enhancement of food intake acoustics from noisy recordings | |
| Strazza et al. | PCG-Decompositor: A New Method for Fetal Phonocardiogram Filtering Based on Wavelet Transform Multi-level Decomposition |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200609 |
|
| RJ01 | Rejection of invention patent application after publication |