CN112580751A - Snore identification device based on ZYNQ and deep learning - Google Patents
Snore identification device based on ZYNQ and deep learning Download PDFInfo
- Publication number
- CN112580751A CN112580751A CN202011616735.9A CN202011616735A CN112580751A CN 112580751 A CN112580751 A CN 112580751A CN 202011616735 A CN202011616735 A CN 202011616735A CN 112580751 A CN112580751 A CN 112580751A
- Authority
- CN
- China
- Prior art keywords
- snore
- module
- data
- neural network
- preprocessing
- 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
- 206010041235 Snoring Diseases 0.000 title claims abstract description 62
- 238000013135 deep learning Methods 0.000 title claims abstract description 13
- 238000007781 pre-processing Methods 0.000 claims abstract description 16
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 13
- 238000004364 calculation method Methods 0.000 claims abstract description 9
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 6
- 238000000034 method Methods 0.000 claims abstract description 5
- 230000008569 process Effects 0.000 claims abstract description 5
- 238000013500 data storage Methods 0.000 claims abstract description 4
- 230000003993 interaction Effects 0.000 claims abstract description 4
- 238000009795 derivation Methods 0.000 claims abstract description 3
- 239000000284 extract Substances 0.000 claims description 9
- 201000002859 sleep apnea Diseases 0.000 claims description 8
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000001914 filtration Methods 0.000 claims description 4
- 238000012549 training Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 238000011176 pooling Methods 0.000 claims description 3
- 230000005236 sound signal Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 4
- 238000002054 transplantation Methods 0.000 abstract 1
- 208000001797 obstructive sleep apnea Diseases 0.000 description 7
- 238000003745 diagnosis Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 208000008784 apnea Diseases 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 230000029058 respiratory gaseous exchange Effects 0.000 description 2
- 208000017667 Chronic Disease Diseases 0.000 description 1
- 206010020591 Hypercapnia Diseases 0.000 description 1
- 206010020772 Hypertension Diseases 0.000 description 1
- 206010021143 Hypoxia Diseases 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 208000026106 cerebrovascular disease Diseases 0.000 description 1
- 208000029078 coronary artery disease Diseases 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000007954 hypoxia Effects 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 230000000422 nocturnal effect Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003860 sleep quality Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
- G06F18/24133—Distances to prototypes
- G06F18/24137—Distances to cluster centroïds
- G06F18/2414—Smoothing the distance, e.g. radial basis function networks [RBFN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR 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
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
- G11B20/10—Digital recording or reproducing
- G11B20/10527—Audio or video recording; Data buffering arrangements
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B31/00—Arrangements for the associated working of recording or reproducing apparatus with related apparatus
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- G—PHYSICS
- G11—INFORMATION STORAGE
- G11B—INFORMATION STORAGE BASED ON RELATIVE MOVEMENT BETWEEN RECORD CARRIER AND TRANSDUCER
- G11B20/00—Signal processing not specific to the method of recording or reproducing; Circuits therefor
- G11B20/10—Digital recording or reproducing
- G11B20/10527—Audio or video recording; Data buffering arrangements
- G11B2020/10537—Audio or video recording
- G11B2020/10546—Audio or video recording specifically adapted for audio data
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computational Linguistics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- General Health & Medical Sciences (AREA)
- Computing Systems (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)
Abstract
The invention discloses a snore identification device based on ZYNQ and deep learning, which comprises a snore acquisition module, an SD card storage module, a snore preprocessing module, a convolutional neural network accelerator IP, a snore judgment module and a conclusion display module. The snore acquisition module is used for acquiring the audio of the detected patient in the sleep state all night; the SD card storage module is used for data interaction in the snore data storage and data calculation processes; the snore preprocessing module is used for preprocessing data before the data enters a network; the general convolutional neural network accelerator IP is used for calculating the derivation part of the Efficient NeT network algorithm; the snore judging module is used for counting based on a network calculation result and identifying the snore according to the AHI; the conclusion display module is used for displaying results; the invention has the advantages of high data processing speed, capability of realizing that portable equipment does not need to use an upper computer and convenience in transplantation and development.
Description
Technical Field
The invention relates to the technical field of snore identification, in particular to a snore identification device based on ZYNQ and deep learning.
Background
Obstructive Sleep Apnea Hypopnea Syndrome (OSAHS) is a chronic disease of Sleep breathing with unknown cause at present, and one of the clinical manifestations of OSAHS is nocturnal Sleep snoring with Apnea. Apnea usually causes hypoxia and hypercapnia, which easily causes a series of complications such as hypertension, coronary heart disease, cerebrovascular disease and the like, so that how to diagnose OSAHS as early as possible is very important.
Based on the above problems, there is a great need for an OSAHS identification device, which has been researched in the related art to identify the OSAHS, and through research and research, most hospitals use a special diagnosis device, namely a Polysomnography (PSG), which is expensive and easily affects sleep, but interferes with the test.
Disclosure of Invention
In order to solve the defects of the prior art, the purposes of reducing test interference and reducing the cost of a test device are realized while the sleep quality is ensured by non-contact measurement, the invention adopts the following technical scheme:
a snore recognition device based on ZYNQ and deep learning comprises: the snore processing system comprises a snore acquisition module, an SD card storage module, a snore preprocessing module, a general convolutional neural network accelerator IP, a snore judging module and a conclusion display module;
the snore acquisition module is used for acquiring the audio of the detected patient in the sleep state all night;
the storage module is used for data interaction in the snore data storage and data calculation processes;
the snore preprocessing module is used for preprocessing data before the data enters a network;
the general convolutional neural network accelerator IP is used for calculating a derivation part of a neural network algorithm;
the snore judging module is used for counting based on a network calculation result and identifying the snore according to an AHI index;
and the conclusion display module is used for displaying results.
Furthermore, the snore collecting module collects snore data of a detected patient by using the microphone array, and in order to enable the data to be effective, the collecting module is placed within 0.5m of the collected patient.
Further, the storage module is an SD card and is used for storing data including a large number of original audio signals, preprocessed signals, and signals during the recognition process.
Furthermore, the snore preprocessing module extracts data from the storage module by a PS (packet switch) end of ZYNQ (zero arrival rate) and filters an original signal, then carries out low-pass filtering on the processed audio by an FIR (finite impulse response) filter, then artificially extracts and marks effective snore section audio from the processed audio, extracts acoustic characteristics from the intercepted effective snore section of the effective signal by an MFCC (Mel frequency cepstrum coefficient) algorithm, caches the extracted effective snore section audio as preprocessed data in a DDR (double data rate), and controls by an AXI-lite bus to wait for DMA (direct memory access) data transmission during subsequent calculation.
Furthermore, the general convolutional neural network accelerator IP is subjected to parameter configuration, data carrying and execution state monitoring by arm, and a convolutional module, a pooling module and a full connection layer module of the deep learning network are all calculated by using the convolutional neural network accelerator IP in the FPGA.
Furthermore, the snore judging module extracts a network tracking training model and judges whether the detected patient has apnea syndrome according to the AHI index.
Furthermore, the conclusion display module selects an OLED display screen using the SPI protocol to display results, so that a user can visually see the results.
The invention has the advantages and beneficial effects that:
the snore identification device integrates acquisition, extraction, training, identification and display, provides an effective way for OSAHS diagnosis, and is more convenient, free of contact and low in cost.
Drawings
FIG. 1 is a block diagram of the apparatus of the present invention.
Detailed Description
The following detailed description of embodiments of the invention refers to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating the present invention, are given by way of illustration and explanation only, not limitation.
As shown in fig. 1, the snore identifying device based on ZYNQ and deep learning network can collect snore information of a detected patient, perform a series of preprocessing in arm of ZYNQ, accelerate the processing of data through a high-efficiency neural network IP, and discriminate and classify the snore, thereby determining the type of apnea syndrome of the patient, including: the snore processing system comprises a snore acquisition module, an SD card storage module, a snore preprocessing module, a general convolutional neural network accelerator IP, a snore judging module and a conclusion display module.
The snore acquisition module acquires snore data of a detected patient by using a microphone array, specifically, a microphone is placed beside the patient according to a specified shape to acquire overnight breathing data, and the data acquired by the module can be instantly stored in an SD card.
The SD card storage module is used for storing a large number of original sound signals, preprocessed signals, signals in the identification process and the like, the storage space of the SD card is preferably configured to be more than 64G, certain transmission speed is required, and data storage and information interaction between the SD card and subsequent modules are facilitated.
The snore preprocessing module is used for preprocessing data before entering a network, and specifically, the preprocessing module operates in ARM of a ZYNQ platform, the device has high requirements on the performance of the platform, the ZYNQ-7000 series and other types with high performance in the same type of platforms are proposed to be used, after the acquisition module finishes acquiring signals all night, the ARM extracts original signals, deburring and filtering the original signals firstly, then a FIR filter is used for low-pass filtering the processed audio, then effective snore section audio is manually extracted and marked from the processed audio, acoustic features are extracted from the intercepted effective snore section of the effective signals through an MFCC algorithm, the extracted effective signals are cached into a DDR and backed up into an SD card as preprocessed data, and the DDR is controlled by an AXI-lite bus to wait for DMA data transmission during subsequent calculation.
The general convolutional neural network accelerator IP is used for hardware acceleration of a deep network, the deep learning network is completed by the cooperation of PS and PL, the PS carries out parameter configuration, carries data and monitors the execution state, a convolutional module, a pooling module and a full connection layer module of the deep learning network are all calculated by using the general convolutional neural network accelerator IP in FPGA, and the obtained characteristic values are also backed up in an SD card so as to be subjected to reverse check and comparison.
The snore judging module is used for judging whether the detected patient has the apnea syndrome, specifically, extracting a characteristic value obtained by a network training model at the PS end of the ZYNQ, judging whether the detected patient has the apnea syndrome according to the AHI index, and synchronously backing up the result into the SD card.
The conclusion display module is used for displaying results, specifically, an OLED display screen using an SPI protocol is selected for use, or other LCDs and TFTLCDs with display functions can be used, the ZYNQ platform is generally provided with an OLED interface, and other display modules are used only by being additionally connected, so that a user can visually see the judgment result of the device.
This device is based on ZYNQ and deep learning, only has better recognition rate to being surveyed patient snore information identification apnea syndrome, collects, draws, trains, discerns, shows as an organic whole, convenient and fast, and whole data all backs up the SD card, can trace to the source and judge, and collection module microphone array and patient contactless do not disturb patient's sleep to whether effective diagnosis is surveyed the patient and has apnea syndrome.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (7)
1. A snore recognition device based on ZYNQ and deep learning comprises: snore collection module, SD card storage module, snore preprocessing module, general convolutional neural network accelerator IP, snore judging module, conclusion display module, its characterized in that:
the snore acquisition module is used for acquiring the audio of the detected patient in the sleep state all night;
the storage module is used for data interaction in the snore data storage and data calculation processes;
the snore preprocessing module is used for preprocessing data before the data enters a network;
the general convolutional neural network accelerator IP is used for calculating a derivation part of a neural network algorithm;
the snore judging module is used for counting based on a network calculation result and identifying the snore according to an AHI index;
and the conclusion display module is used for displaying results.
2. The device of claim 1, wherein the snore collecting module collects snore data of a patient to be detected by using a microphone array, and the collecting module is placed within 0.5m of the person to be detected.
3. The device of claim 1, wherein the memory module is an SD card for storing data including a plurality of original sound signals, preprocessed signals, and signals during recognition.
4. The device for identifying snore based on ZYNQ and deep learning as claimed in claim 1, wherein the snore preprocessing module extracts data from the storage module by the PS end of ZYNQ, firstly removes the burr of the original signal and filters, then uses FIR filter to perform low pass filtering on the processed audio, then manually extracts and marks the effective snore section audio from the processed audio, extracts the acoustic characteristics of the effective signal from the intercepted effective snore section by MFCC algorithm, caches the extracted effective signal as the preprocessed data in DDR, is controlled by AXI-lite bus, and waits for DMA data transmission during subsequent calculation.
5. The device of claim 1, wherein the general convolutional neural network accelerator IP is configured with parameters, carries data, and monitors execution status by arm, and the convolutional module, pooling module, and full connection layer module of the deep learning network are all calculated by using the convolutional neural network accelerator IP in FPGA.
6. The device of claim 1, wherein the snore determining module extracts a network tracking training model and determines whether the detected patient has apnea syndrome according to an AHI index.
7. The device of claim 1, wherein the conclusion display module selects an OLED display screen using SPI protocol for displaying the result, so that the user can visually see the result.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011616735.9A CN112580751A (en) | 2020-12-31 | 2020-12-31 | Snore identification device based on ZYNQ and deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011616735.9A CN112580751A (en) | 2020-12-31 | 2020-12-31 | Snore identification device based on ZYNQ and deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112580751A true CN112580751A (en) | 2021-03-30 |
Family
ID=75144885
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011616735.9A Pending CN112580751A (en) | 2020-12-31 | 2020-12-31 | Snore identification device based on ZYNQ and deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112580751A (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110013222A (en) * | 2019-04-03 | 2019-07-16 | 杭州电子科技大学 | A kind of system for sleep apnea detection |
| CN110414401A (en) * | 2019-07-22 | 2019-11-05 | 杭州电子科技大学 | A kind of intelligent monitor system and monitoring method based on PYNQ |
| CN111613210A (en) * | 2020-07-06 | 2020-09-01 | 杭州电子科技大学 | Categorised detecting system of all kinds of apnea syndromes |
-
2020
- 2020-12-31 CN CN202011616735.9A patent/CN112580751A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110013222A (en) * | 2019-04-03 | 2019-07-16 | 杭州电子科技大学 | A kind of system for sleep apnea detection |
| CN110414401A (en) * | 2019-07-22 | 2019-11-05 | 杭州电子科技大学 | A kind of intelligent monitor system and monitoring method based on PYNQ |
| CN111613210A (en) * | 2020-07-06 | 2020-09-01 | 杭州电子科技大学 | Categorised detecting system of all kinds of apnea syndromes |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108670200B (en) | Sleep snore classification detection method and system based on deep learning | |
| CN109009129B (en) | Sow respiratory disease early warning system based on acoustic analysis | |
| CN103558029B (en) | A kind of engine abnormal noise on-line fault diagnosis system and diagnostic method | |
| CN110160765B (en) | Impact characteristic identification method and system based on sound or vibration signal | |
| CN109480783B (en) | Apnea detection method and device and computing equipment | |
| CN107080545B (en) | A kind of lie detection system based on brain electricity | |
| WO2021114761A1 (en) | Lung rale artificial intelligence real-time classification method, system and device of electronic stethoscope, and readable storage medium | |
| CN111248859A (en) | Automatic sleep apnea detection method based on convolutional neural network | |
| CN107105092A (en) | A kind of human body tumble recognition methods based on dynamic time warping | |
| CN111544006B (en) | Wearable equipment for quantifying and identifying dyskinesia of Parkinson's disease people | |
| CN111613210A (en) | Categorised detecting system of all kinds of apnea syndromes | |
| CN111783715A (en) | Identity recognition method based on pulse signal feature extraction | |
| CN204744109U (en) | Cough judgement system based on accelerometer | |
| CN112580751A (en) | Snore identification device based on ZYNQ and deep learning | |
| CN106539560A (en) | Sound of snoring detection method, sound of snoring detecting system, snoring system and Easy pillow | |
| CN201683910U (en) | Intelligent cardiopulmonary analyzing instrument | |
| CN117064333B (en) | Primary screening device for obstructive sleep apnea hypopnea syndrome | |
| CN111048089B (en) | Method for improving voice wake-up success rate of intelligent wearable device, electronic device and computer readable storage medium | |
| CN106419884A (en) | Heart rate calculating method and system based on wavelet analysis | |
| CN104155966B (en) | Automobile CDC data acquisition analysis apparatus | |
| KR102536544B1 (en) | Wearable device for monitoring heart disease based on combined measurement of heart sound and electrocardiogram and operation method thereof | |
| CN114420304A (en) | Novel new crown auxiliary screening method and device based on deep learning | |
| CN113509169A (en) | Multi-parameter-based non-contact sleep apnea detection system and method | |
| CN111131950B (en) | Method for controlling microphone switch through triaxial acceleration sensor, electronic equipment and computer readable storage medium | |
| CN109961090B (en) | Behavior classification method and system based on intelligent wearable device |
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 |