CN102697520B - Electronic stethoscope based on intelligent distinguishing function - Google Patents
Electronic stethoscope based on intelligent distinguishing function Download PDFInfo
- Publication number
- CN102697520B CN102697520B CN201210140040.7A CN201210140040A CN102697520B CN 102697520 B CN102697520 B CN 102697520B CN 201210140040 A CN201210140040 A CN 201210140040A CN 102697520 B CN102697520 B CN 102697520B
- Authority
- CN
- China
- Prior art keywords
- signal
- sounds
- unit
- hear sounds
- processor unit
- 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.)
- Expired - Fee Related
Links
Classifications
-
- G06F19/00—
Abstract
The invention belongs to the technical field of medical machinery, and relates to an electronic stethoscope which is used for auxiliary diagnosis and is capable of intelligently distinguishing physiological parameters, such as heart sounds and breath sounds, judging the types of the heart sounds and the breath sounds, and extracting the disease characteristics. The electronic stethoscope comprises a processor unit, and a signal collection unit, an external drive unit, a storage unit and a data bus interface unit which are connected with the processor unit, wherein the signal collection unit is used for collecting the signals of the heart sounds and the breath sounds and implementing the pre-processing on the signals; the processor unit is used for implementing the pattern recognition algorithms on the heart sounds and the breath sounds, separating the two sounds from each other, intelligently distinguishing and classifying the signals of the heart sounds and the breath sounds, and managing the other hardware units; the storage unit is used for storing the programs and the extension programs thereof, storing the data of the heart sounds and the breath sounds, the standard heart sounds and the breath sounds, and the audition model of each typical case of the disease, and outputting and playing; and the external drive unit and the data bus interface unit are used for implementing external operation function drive and data communication.
Description
Technical field
The invention belongs to technical field of medical instruments, relate in particular to the type of the physiological parameters such as a kind of Intelligent Recognition hear sounds for auxiliary diagnosis and respiratory murmur, differentiation hear sounds and respiratory murmur, the electronic stethoscope of extraction genius morbi.
Background technology
Hear sounds and respiratory murmur are in human heart motion and the physiological feature producing in human body respiration motion, and it is containing physiology and the pathological information of heart and respiratory system.In clinical medicine the pathological analysis of hear sounds and respiratory murmur and vim and vigour just the technology such as assay, X-ray line, B ultrasonic diagnosis compare, more conveniently early prediction heart and respiratory system disease.Explore and study the technical know-how very long time of existing 120 years of the clinical diagnosis of hear sounds and respiratory murmur, formed and comparatively improved auscultation diagnosis subject system extensive use in clinical practice and teaching practice.Simultaneously along with the development of science and technology application, stethoscope also develops into electronic type at development and improvement constantly from mechanical type, is all to help doctor more can accurately catch physiology and pathological information.But the limitation due to the function of our people's ear auditions own, to be difficult to realize for hear sounds and respiratory murmur auscultation effect accurately, and people produce with the research of the aspects such as mechanism of propagating not deep enough to the acoustic information of hear sounds and respiratory murmur, for need deep research and the grasp of dependency between acoustic information and pathology, physiology, these have all affected current clinical auscultation diagnostics's application and development.
Stethoscope is the just innovation in hardware link in the evolution to electronic type, form is the filtering accepting sound and carry out various electronic hardware again of electronics mike, operation amplifier, storage etc., realizes the function of acoustic information recording generating digital file, storage and the repeat playing of auscultation.These objects be all allow the hear sounds heard and respiratory murmur signal more clear accurately, but due to overlapping on frequency spectrum of the sound interference noise of hear sounds and respiratory murmur and signal, the general hardware capability of simple employing often can not be eliminated interference noise effectively, need to adopt more effective signal processing technology to eliminate these interference noises, so current electronic stethoscope can't truly assist doctor to carry out clinical diagnosis.
Therefore clinically high in the urgent need to a kind of accuracy, be easy to handle, the auxiliary doctor's of Intelligent Recognition of easy to carry, quick diagnosis auscultation apparatus.Allow clinician's apparatus in hear sounds and respiratory auscultation go out to hear at that time that by the science recognition methods Intelligent Recognition having possessed the feature of sound judges classification in unison, and make and analyze make a comment or criticism out the type of disease and the pathological information containing, effectively utilize auscultation technology to realize the auxiliary diagnosis of early discovery disease sign.The research and development that the mature modern signal processing of industry and mode identification technology are this kind of auxiliary diagnosis equipment are at present laid a good foundation, the electronic stethoscope of the Intelligent Recognition that the most advanced microelectronic modules equipment that lot of domestic and foreign manufacturer provides is new generation has been created condition, and application embedded electronic information technology and sensor technology have formed the technological means of research and development.
Summary of the invention
The present invention for solve the technical problem existing in known technology provide a kind of accuracy high, be easy to handle, the electronic stethoscope based on intelligent recognition function of easy to carry, quick diagnosis.
The technical scheme that the present invention takes for the technical problem existing in solution known technology is: the electronic stethoscope based on intelligent recognition function comprises:
Processor unit, for specific implementation to the noise of hear sounds and respiratory murmur signal separate out, extract hear sounds and respiratory murmur case feature, extract hear sounds and respiratory murmur case characteristic signal envelope characteristic, realize automatic segmentation algorithm identified and neural computing; Signal gathering unit, is connected with processor unit, for gather hear sounds respiratory murmur signal and to its amplify, filtering, etc. processing the analogue signal collecting is converted to digital signal; Peripheral driver unit, is connected with processor unit, for each operating function link is driven as keyboard, touch screen etc. provides to control; Memory element, is connected with processor unit, for program and extender thereof are stored, also for storing the audition pattern of each model case of the hear sounds of hear sounds respiratory murmur data and standard and respiratory murmur, disease; Data bus interface unit, is connected with processor unit, for the connection of 4G/3G module, WiFi module, GPRS module, WLAN module and LAN module.
The present invention has adopted following technical scheme:
The dual processor of described processor unit for being formed by embedded microprocessor ARM and DSP.
Described signal gathering unit comprises sensor, signal conditioning circuit and audio chip; Described sensor consists of traditional stethoscope tin, hose and Miniature ECM; Described Miniature ECM is positioned at one end of hose, and its signal lead-out wire is connected to the input of described signal conditioning circuit; Described signal conditioning circuit comprises amplifying circuit and filter circuit; Described audio chip is located between described signal conditioning circuit and described processor unit, for realizing collection and the playback of audio signal, also for playing the audition pattern of the default standard hear sounds of memory element, respiratory murmur and each model case of disease.
Also comprise application performance for improving instrument, with the data communication units that other equipment communicate, described data communication units is the serial communication interface that is connected to described processor unit.
Described data communication units also comprise for realize with PC carry out data interaction usb interface unit, described usb interface unit is directly connected with described processor unit.
Described memory element comprises for storing the FLASH memorizer of data and the SDRAM memorizer of carrying out for extender, described FLASH memorizer and SDRAM memorizer are all connected directly to described processor unit, and be also connected to each other between the two, described SDRAM memorizer reads and writes the data in FLASH memorizer when extender is carried out.
Described FLASH memorizer is also connected to described processor unit simultaneously, and is also provided with between the two and for extensible processor, described FLASH memorizer is carried out the latch FPGA of high bit addressing.
Described memory element also comprises the SDCARD that is connected directly to described processor unit, carries out the storage of hear sounds respiratory murmur data as main storage.
Also comprise the diamagnetic circuit unit for shield electromagnetic interference.
Also comprise power subsystem, described power subsystem comprises polymer Li-ion battery.
The advantage that the present invention has and innovation effect are: modern signal processing and mode identification technology have been incorporated in electronic stethoscope, and application embedded electronic information technology and sensor technology are realized the electronic stethoscope based on intelligent recognition function.The algorithm for pattern recognition of the processor unit specific implementation being formed by ARM+DSP dual micro processor to hear sounds and respiratory murmur, complete Intelligent Recognition and classification to hear sounds respiratory murmur signal, memory element is not only for storing current gained hear sounds respiratory murmur data, and be preset with hear sounds, the standard audition pattern of each model case of respiratory murmur disease, by audio chip, the audio signal collecting carried out to playback and can select to play and listen to above-mentioned standard hear sounds, the audition pattern of respiratory murmur and each model case of disease, be user-friendly for contrast judgement, and the accuracy of enhancing judgement, reduce the impact of subjective factors as far as possible.Serial communication interface in data communication units and USB interface can communicate with other equipment instrument, the data interaction especially and between PC; The data bus interface unit that connects 4G/3G module, WiFi module, GPRS module, WLAN module and LAN module allows instrument realize and the data interaction of network environment and the data communication of Internet, coordinates network assistance diagnostic system more can bring into play its advantage.
Accompanying drawing explanation
Fig. 1 is system block diagram of the present invention;
Fig. 2 is the structural representation of Sensor section in Fig. 1.
The specific embodiment
For further understanding summary of the invention of the present invention, Characteristic, hereby exemplify following examples, and coordinate accompanying drawing to be described in detail as follows:
Fig. 1 is system block diagram of the present invention.Electronic stethoscope based on intelligent recognition function comprises processor unit, for specific implementation to the noise of hear sounds and respiratory murmur signal separate out, extract hear sounds and respiratory murmur case feature, extract hear sounds and respiratory murmur case characteristic signal envelope characteristic, realize automatic segmentation algorithm identified and neural computing.Signal gathering unit, is connected with processor unit, for gather hear sounds respiratory murmur signal and to its amplify, filtering, etc. processing the analogue signal collecting is converted to digital signal.Peripheral driver unit, is connected with processor unit, for each operating function link is driven as keyboard, touch screen etc. provides to control.Memory element, is connected with processor unit, for program and extender thereof are stored, also for storing the audition pattern of each model case of the hear sounds of hear sounds respiratory murmur data and standard and respiratory murmur, disease.Data bus interface unit, is connected with processor unit, for the connection of 4G/3G module, GPRS module, WLAN module and LAN module.
The stethoscopic embedded microprocessor of Intelligent Recognition of the present invention, application ARM+DSP center processing unit C6A816x, integrated ARM CortexA9 kernel and single kernel floating-point and the fixed DSP with 1.5GHz, with this, realize the general-purpose algorithm of modern signal processing and mode identification technology, as data filtering, image are processed, intelligent identification technology algorithm, innovation realization is applied MATLAB system emulation and has been realized complicated mathematical function computing, Time-Frequency Analysis in this core processor unit, the Processing Algorithm of wavelet conversion, neural computing and mode identificating software function.Adopt just ARM CortexA9 to realize that graphic user interface (GUI), Communication processing are connected with network, system is controlled and the processing application of concrete operations function, the high-performance embedded hardware structure of the high throughput of its double-core and DSP, to ensure the realization of Intelligent Recognition algorithm function.
The present invention realizes hear sounds and the automatic identification and analysis of respiratory murmur signal, its recognition methods adopting is: the time-frequency domain characteristic of utilizing hear sounds and each composition of respiratory murmur, de-noising is carried out in the reconstruct of application wavelet decomposition, correct baseline drift, then extract its envelope signal, and then from envelope signal, extract the characteristic information of hear sounds and each composition of respiratory murmur, in conjunction with the relevant medical knowledge of hear sounds and respiratory murmur, utilize the method for neutral net hear sounds is trained and identify, make doctor easy to use.
Fig. 2 is the structural representation of Sensor section.Signal gathering unit comprises sensor, signal conditioning circuit and audio chip.Wherein, sensor consists of traditional stethoscope tin 1, hose 2 and Miniature ECM 3, and Miniature ECM 3 is positioned at one end of hose 2, and its signal lead-out wire 10 is connected to the input of signal conditioning circuit.Signal conditioning circuit comprises amplifying circuit and filter circuit.Audio chip selects TLV320AIC3253 to be located between signal conditioning circuit and processor unit, for realizing collection and the playback of audio signal, also for playing the audition pattern of the default standard hear sounds of memory element, respiratory murmur and each model case of disease.
Audio chip TLV320AIC3253 carries out modulus (A/D) conversion to hear sounds respiratory murmur signal.The abundant stereo coding/decoding sound function of applying high-performance, being integrated with analog functuion, built-in Mike and earphone out amplifier, support MIC and two kinds of input modes of LINE IN, and input and output are all had to programmable-gain adjusting.The analog digital conversion of AIC3253 (A/D) and digital-to-analogue conversion (D/A) component height are integrated in chip internal, the integrated digital interpolation wave filter of high sampling rate in inside.Owing to having above-mentioned advantage, making AIC3253 is a ideal audio frequency simulation I/O device, can be applied in well various digital audio fields, particularly applicable for this class physiology acoustical signal of hear sounds respiratory murmur.
Applying suitable sensor acquisition signal is the basis of accurately collecting hear sounds respiratory murmur signal, because hear sounds respiratory murmur signal is fainter, this has proposed very high requirement to the sensitivity of sensor, definition, precision and stability, necessary selection high configuration, miniaturization and, the sensor that signal to noise ratio is high and easy operating is used.Should be convenient to and human body skin Surface Contact, reduce again the impact on human body surface vibration and environment noise as far as possible.In the prior art, the satisfactory PVDF of having piezoelectric transducer, condenser microphone and add Air Coupling chamber condenser microphone, electret capcitor microphone.This equipment choosing Miniature ECM.
The realization of signal conditioning circuit: be directly very faint from the hear sounds respiratory murmur signal of sensor output, sensor is placed to the 4th root bone under human body left collarbone vacancy time record signal peak-to-peak value be 20mV, this signal is mainly cardiechema signals, the respiratory murmur signal and the noise that comprise part; Under the clavicle of right side, to record the peak-to-peak value of signal be 5mV in the vacancy of the second root bone, and this signal is mainly respiratory murmur signal, comprises part hear sounds and noise; At sensor, not contacting the peak-to-peak value that records noise under people's concrete conditions in the establishment of a specific crime is 2mV.The faint ac small signal that this class comprises some strength noise is unfavorable for the collection of data, must amplify it, a series of conditionings such as filtering, is provided with thus signal conditioning circuit.
Operation amplifier adopts AD827 to carry out processing and amplifying to the signal of sensor output.AD827 is the operational amplifier of a high-performance low-noise, has good noiseproof feature, and input noise is only 5nV, has good amplifying power and quite high small signal bandwidth simultaneously.
The hear sounds respiratory murmur signal detecting through sensor is filter away high frequency noise signal first, then through AD827, amplifies.For the better higher noise of rejection frequency, by adding filter capacitor to reduce the high-frequency gain of amplifying circuit in the feedback circuit of operational amplifier.In actual realization, when gain is 10dB, under human body left collarbone, the peak-to-peak value of the vacancy output signal of the 3rd to four root bones is 200-300mV; Under the clavicle of human body right side, the peak-to-peak value of the vacancy output signal of the second root bone is 150mV; The peak-to-peak value that does not contact noise under people's concrete conditions in the establishment of a specific crime at sensor is 10mV.The hear sounds and the respiratory murmur signal that amplify can meet system requirements.Because the main frequency scope of hear sounds is 20Hz-300Hz, the main frequency scope of respiratory murmur is 25Hz-2000Hz, in order better to suppress the out-of-band noise of this scope, system has the low pass filter that a cut-off frequency is 2000Hz to carry out filtering additional noise at the outfan of amplifier.
After voice signal input, pretreatment and digitized are the preconditions of carrying out speech recognition.Wherein, pretreatment is mainly to carry out pre-filtering, retains the voice signal of 20～3400Hz of normal person; Digitized is to carry out A/D conversion and the processing such as anti-aliasing; Feature extraction is to carry out voice signal training and identify requisite step.The parameter that can embody phonic signal character comprises: the cepstrum parameter based on LPC; The cepstrum parameter of Mel coefficient; Adopt the feature analysis means of forward position Digital Signal Processing, as wavelet analysis, time/frequency-domain analysis, artificial neural network etc.First adopt the cepstrum parameter method for expressing based on LPC, the eigenvalue extracting deposits in reference model storehouse, and then with the cepstrum parameter of finding Mel coefficient, is used for mating the eigenvalue of voice signal to be identified.It is the core of carrying out speech recognition that coupling is calculated, by sample voice value to be identified after feature extraction, the template producing during with systematic training is mated, and in identification identification, gets with the corresponding voice of model of voice similarity maximum to be identified as recognition result.
The realization of memory circuit: at the inner integrated RAM of certain capacity of DSP digital signal processor, but the driver that relates to adaptive filter algorithm, wavelet algorithm and various peripheral hardwares due to native system is larger, internal RAM cannot meet the requirement of working procedure, therefore need to be at DSP sheet external expansion program storage.SDRAM is synchronous DRAM, need to constantly refresh and could keep data, and rank addresses is multiplexing, and read or write speed is slower, but capacity is larger, and cost is lower.The EMIF interface of DSP digital signal processor is through connecting polytype SDRAM after configuration.For reserved larger memory space, adapt to the needs of complicated algorithm, be originally embodied as and select SDRAM as main storage.
Program and data are moved in SDRAM, but work as system power failure, and the program in SDRAM and data will disappear, and therefore, at this moment just a kind of nonvolatile memory of needs is used for save data.This conventional class memorizer comprises: FLASH, disk, various storage cards etc., native system selects FLASH as program storage, the FLASH chip Am29F010B that has adopted AMD to produce.This chip has that capacity is large, volume is little, low in energy consumption, low cost and other advantages.Between FLASH memorizer and above-mentioned SDRAM memorizer, be also connected to each other, SDRAM memorizer reads and writes the data in FLASH memorizer when extender is carried out.
Data communication realizes: in order to improve the application performance of system, combine with some current remote medical diagnosis systems, serial communication interface in data communication units and USB interface can communicate with other equipment instrument, the data interaction especially and between PC; The data bus interface unit that connects 4G/3G module, GPRS module, WLAN module and LAN module allows instrument realize and the data interaction of network environment and the data communication of Internet, coordinates network assistance diagnostic system more can bring into play its advantage.
SDCARD Interface realization is: in order there to be enough space storage hear sounds respiratory murmur signals, native system adopts SD card as hear sounds respiratory murmur signal main storage device.An integrated SD card control module in dsp chip, communication is convenient, speed is fast.
Native system is provided for the diamagnetic circuit unit of shield electromagnetic interference.
Native system arranges power subsystem, for whole system provides electric power support, adopts polymer Li-ion battery in power subsystem, and capacity is at 1500-2000mAH.
Electronic stethoscope of the present invention be take dual micro processor as core, on the basis of above-mentioned hardware structure, by embedded software, is researched and developed and is realized concrete function.Specific implementation process is as follows:
The DSP embedded program design of realization based on MATLAB software platform, make full use of the power of MATLAB science software for calculation, with configuration type modeling graphical, based on model, the emulation mode combining in conjunction with software emulation and software and hardware is verified the correctness of algorithm, generate quickly and efficiently and run directly in the embedded real-time C code stand-alone program on target developing plate, the program of moving on debug target plate in real time by DSP development environment under MATLAB environment and change relevant parameter.The quickness and high efficiency that fully application code generates, the upgrading of software system and be easy to safeguard for the implant system of hardware platform.
Embedded MATLAB (Embedded MATLAB) in application MATLAB software, Simulink, RTW (RTW Embedded Coder), the dsp software system synthesis development platform that the IDE (as the Code Composer Studio IDE for TI) of Embedded IDE Link and Target Suppoa Package and exploitation DSP application program builds, the embedding people formula application software design concept of employing based on model, comprise algorithm design and emulation, code and project generate, code verification and on-line debugging etc. are in interior system research and development task.
On MATLAB platform, design the step of DSP embedded real-time application, with algorithm model, start, add and configure target selection module (Target References block), simulation parameter is set, compilation model, the code that operation generates on processor.From algorithm design to the whole link of verifying, realizing from concept to code, all emphasis drop in the checking and algorithm development of scheme, after code development T.G Grammar, can be directly embedded into chip.
The concrete step realizing in the present invention:
(1) motion control arithmetic design and emulation: the modules that calls the needed chip of corresponding dsp chip in Matlab/Simulink, in conjunction with relevant control module in Simulink, build system control model, utilize the copying that Simulink is powerful, the system control model of putting up is carried out to real-time simulation, regulate the corresponding parameter of controlling, until reach control requirement in emulation; Use Matlab software, the result to the workbox of Simulink to the output of ARM+DSP primary module analog simulation, the effect of viewing system control module is to improve the efficiency of design;
(2) automatic code conversion: the code that utilizes automatic code systematic function that the good control model conversion of emulation is become to need, set the type (as: assembly code of wanting generating code simultaneously, C code or C++ code), generate corresponding control routine, and automatically code is forwarded in CCS development environment;
This step is utilized automatic code systematic function, makes developer in the situation that not writing line code, obtain complicated motion control code, has reduced the exploitation threshold of motion control arithmetic exploitation, has improved the development efficiency of motion control arithmetic program;
(3) code is transplanted: motion control code is carried out in CCS development environment comprehensive, be transplanted in embedded real-time operating system after making necessary modification;
Adopt embedded development method, bring very large facility to the transplanting of code, this step is carried out the programing work of assembly code or C code under CCS development environment, automatically the code generating carries out necessary modification, the needs that meet concrete Application and Development, finally have transplanting the embedded real time system code compilation of motion control arithmetic to become final executable code;
(4) checking: the binary code generating in CCS environment, by SEEDXDS-usb2.0 simulation programmer, is downloaded in the processor (DSP) of universal motion controller, carry out the online motor control of practical object, and carry out data record work;
This step is used SEEDXDS-usb2.0 simulation programmer, correct by the loading of executed file of compiling in processor, in hardware entities, be to debug, carry out data record, to again revise;
(5) information feedback: the control inputs of the data of record and expectation is contrasted, comparing result is fed back in the control model in Simulink, if actual result and expected value disagree, accordingly the system control model in Simulink or systematic parameter are modified;
This step provides test data for the effective optimization of Intelligent Recognition algorithm, programmes recording on the basis of actual result again, has improved the efficiency of application algorithm development;
(6) repeat the parameters of above step adjustment control, improve the algorithmic procedure of whole Intelligent Recognition.
The present invention can be according to algorithm routine the needs of the ruuning situation on hardware platform and products-hardware system, on the basis of hardware platform, the synchronous design of carrying out algorithm routine and products-hardware circuit, to have improved development efficiency; Adopt the automatic code dress of Matlab to change technology, can well convert the various Intelligent Recognition algorithms that are easy to realize to the readable C code of writing in Matlab, and can be good at transplanting to be embedded in CCS development environment calling, the C code of these motor controles is very complicated, often need professional experienced programmer to write, so just reduced greatly the threshold of northern system development, and cycle and the cost of exploitation are reduced.
The specific implementation of the intelligent identification Method in the present invention is:
One, the noise of heart respiratory murmur signal is separated out.In gathering the process of heart respiratory murmur signal, unavoidably have noise and mix wherein, these noises not only affect the accuracy of acquired signal, may override the effective ingredient of measured signal in more serious situation completely.Thereby before heart respiratory murmur being carried out to careful analysis identification, must first to the signal collecting, carry out pretreatment, manage filtering and weaken noise.These noises comprise particularly:
(1) random noise
In order to receive weak heart respiratory murmur signal, the sensitivity of microphone of generally selecting is all very high, during auscultation when surrounding has larger noise jamming, or when stethoscope directly contacts with human body and produces friction noise, will be received by microphone simultaneously, amplify and enter system, making to mix into interfering signal in useful signal.
(2) getting rid of power frequency disturbs
There is electromagenetic wave radiation in measurement environment electrical equipment around, this electromagnetic interference be take 50Hz power frequency as main component, and electronic stethoscope self circuit also may be subject to power frequency interference effect in power supply, and the power frequency of adulterating when processing signals is disturbed, and makes signal band jagged.
(3) lowering apparatus self-noise
Due to the defect of electronic machine itself, will produce such as the discrete noise of electronic device, the thermal noise of resistance etc.These noises and measuring-signal are mixed in together, and instrument is declined to the resolution capability of measuring-signal.In addition due to the complexity of physiological signal, in the gatherer process of heart respiratory murmur signal, also can be subject to interference and the impact of many other physiological signals, this problem also can not be ignored.For these noises, can reduce impact by improving the performance of electronic stethoscope, as adopted shielding, filtering technique etc.
Two, wavelet transformation hear sounds and the feature extraction of respiratory murmur case.
Time frequency window in wavelet transformation has adaptivity, so can obtain more reflecting the information of hear sounds and respiratory murmur case feature.Wavelet transformation carries out automatic segmentation and extracts characteristic parameter the hear sounds of two time periods in a window and respiratory murmur signal, has realized two analysis purposes in this window, and one is the segmentation of X1 and X2, and another is feature extraction.After segmentation, with wavelet transformation, decompose hear sounds and respiratory murmur signal, and the details of being decomposed by layer 6 (high frequency) coefficient obtains characteristic vector, thereby use dynamic design analysis to obtain optimum characteristic parameter.
Three, the envelop feature extraction of hear sounds and respiratory murmur case distinctive tone signal.
Envelope extraction algorithm is to adopt after the signal envelope that Hilbert transform obtains, repeatedly find the maximum point of new envelope, with three times, than Splines Interpolation Curve matching, also connect in turn these maximum points, obtain repeatedly various envelope, be finally defined as approaching pattern envelope.
Four, automatic segmentation algorithm identified, after pretreatment and envelope extraction above, what hear sounds and respiratory murmur signal location can be obvious is outstanding, and the realization detecting for hear sounds and respiratory murmur automatic segmentation provides may.But because physiological signal exists randomness and mutability, widely different between the features such as the duration that actual hear sounds and each composition of respiratory murmur signal show, amplitude, frequency, therefore to the signal envelope having obtained, also must coordinate a series of suitable detection of dynamic strategies, eliminate the false and retain the true, just can reach the requirement of accurate automatic segmentation.Through constantly test adjustment, concrete steps are:
(1) according to the signal envelope of Hilbert transform and the acquisition of cubic spline difference curve, from the peak value of envelope, calculate two parameters as threshold value, because the amplitude of heart respiratory murmur signal is fainter, threshold value value is wanted suitable stablizing;
(2) obtain all limit codomains of envelope, give up the excessive minimum point of those peak values;
(3) envelope of signal is extracted in application Hilbert transform, original real signal is corresponded to a kind of mapping of complex space.For a real signal X (t), the mathematic(al) representation of its analytic signal is as follows: Z (t)=X (t)+Y v (t), the real part in formula and imaginary part are the real function that meets respectively Cauchy-Riemann equations.The imaginary part Yv of analytic signal (t), can try to achieve hear sounds and respiratory murmur signal have been carried out to effective pretreatment by real signal X (t) being carried out to Xi Er baud conversion, the basic feature that has retained hear sounds and respiratory murmur, for the segmentation of various complicated hear sounds and respiratory murmur is laid a good foundation.
Five, neutral net eigenvalue chooses.
The whole effective informations that comprised signal in time domain waveform characteristic vector time domain waveform, the little and length of deterministic signal or data variation is at the stochastic signal of certain limit, and total data is as the input feature vector value of neutral net.Hear sounds respiratory murmur is only done after simple process, its time domain data, directly as input feature vector vector, is kept to the objective reality of original signal, give full play to the recognition function of neutral net self.Should be in this way in the present invention the main process chip ARM+DSP of CPU can adapt to the high ability enhance operation speed of calculating of Complex Neural Network in large scale, realize first the stochastic signal that hear sounds respiratory murmur signal belongs to nonlinear and nonstationary, directly the realization of the input feature vector using it as neutral net.
Realizing artificial neural network is the system that structurally mimic biology nerve connects, and carries out pattern recognition work.Application nerve network system need to have the ability of storage information to train system, and system can provide the output of expectation after training for any given input.
Claims (10)
1. the electronic stethoscope based on intelligent recognition function, is characterized in that: comprises,
Processor unit, for specific implementation, to processing, the noise of hear sounds and respiratory murmur signal, separate out, classify extractions, standard pattern contrasts, by being embedded in the Intelligent Recognition algorithm extraction hear sounds of this element and the envelope characteristic of respiratory murmur case feature, identification hear sounds and respiratory murmur case characteristic signal, realizing segmentation algorithm identification and neural computing, realize the Intelligent Recognition to hear sounds and respiratory murmur signal;
Signal gathering unit, is connected with processor unit, for gather hear sounds Lung Sounds and to its amplify, Filtering Processing the analogue signal collecting is converted to digital signal;
Peripheral driver unit, is connected with processor unit, for each operating function unit being provided control, drives;
Memory element, is connected with processor unit, for program and extender thereof are stored, also for storing the audition pattern of each model case of the hear sounds of hear sounds respiratory murmur and standard and respiratory murmur, disease;
Data bus interface unit, is connected with processor unit, for 4G the connection of 3G module, WiFi module, GPRS module, WLAN module and LAN module.
2. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: the dual processor of described processor unit for being formed by embedded microprocessor ARM and DSP.
3. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: described signal gathering unit comprises sensor, signal conditioning circuit and audio chip; Described sensor consists of traditional stethoscope tin, hose and Miniature ECM; Described Miniature ECM is positioned at one end of hose, and its signal lead-out wire is connected to the input of described signal conditioning circuit; Described signal conditioning circuit comprises amplifying circuit and filter circuit; Described audio chip is located between described signal conditioning circuit and described processor unit, for realizing collection and the playback of audio signal, also for playing the audition pattern of the default standard hear sounds of memory element, respiratory murmur and each model case of disease.
4. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: also comprise data communication units, this data communication units comprise 4G 3G module, WiFi module, GPRS module, WLAN module and LAN module, in order to the data communication between realization and the Internet and master-slave equipment.
5. according to the electronic stethoscope based on intelligent recognition function claimed in claim 4, it is characterized in that: described data communication units also comprises the usb interface unit that carries out data interaction with PC for realizing, described usb interface unit is directly connected with described processor unit.
6. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: described memory element comprises for storing the FLASH memorizer of data and the SDRAM memorizer of carrying out for extender, described FLASH memorizer and SDRAM memorizer are all connected directly to described processor unit, and be also connected to each other between the two, described SDRAM memorizer reads and writes the data in FLASH memorizer when extender is carried out.
7. according to the electronic stethoscope based on intelligent recognition function claimed in claim 6, it is characterized in that: described FLASH memorizer is also connected to described processor unit simultaneously, and be also provided with between the two and for extensible processor, described FLASH memorizer carried out the latch FPGA of high bit addressing.
8. according to the electronic stethoscope based on intelligent recognition function described in claim 6 or 7, it is characterized in that: described memory element also comprises the SDCARD that is connected directly to described processor unit, carries out the storage of hear sounds respiratory murmur data as main storage.
9. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: also comprise the diamagnetic circuit unit for shield electromagnetic interference.
10. according to the electronic stethoscope based on intelligent recognition function claimed in claim 1, it is characterized in that: also comprise power subsystem, described power subsystem comprises polymer Li-ion battery.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210140040.7A CN102697520B (en) | 2012-05-08 | 2012-05-08 | Electronic stethoscope based on intelligent distinguishing function |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210140040.7A CN102697520B (en) | 2012-05-08 | 2012-05-08 | Electronic stethoscope based on intelligent distinguishing function |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102697520A CN102697520A (en) | 2012-10-03 |
| CN102697520B true CN102697520B (en) | 2014-10-29 |
Family
ID=46890791
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210140040.7A Expired - Fee Related CN102697520B (en) | 2012-05-08 | 2012-05-08 | Electronic stethoscope based on intelligent distinguishing function |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102697520B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3100675A1 (en) * | 2015-06-03 | 2016-12-07 | IMEDI PLUS Inc. | Method and system for recognizing physiological sound |
Families Citing this family (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201618720A (en) | 2014-11-20 | 2016-06-01 | 創心醫電股份有限公司 | Method, organ sound recorder, and system for organ sound acquisution |
| CN104523289A (en) * | 2015-01-27 | 2015-04-22 | 河北联合大学 | Lung lobe breath sound monitoring and automatic analyzing device |
| CN104887263B (en) * | 2015-05-21 | 2018-09-07 | 东南大学 | A kind of identification algorithm and its system based on heart sound multi-dimension feature extraction |
| CN106308845A (en) * | 2015-06-23 | 2017-01-11 | 黄楚 | Intelligent remote stethoscope and using method thereof |
| CN104970820A (en) * | 2015-06-26 | 2015-10-14 | 迈德高武汉生物医学信息科技有限公司 | Intelligent auscultation system and detecting method |
| CN105943077A (en) * | 2015-09-29 | 2016-09-21 | 刘伟锋 | Stethoscope |
| CN105496447B (en) * | 2016-01-15 | 2019-02-05 | 厦门大学 | Electronic auscultation device with active noise reduction and auxiliary diagnosis function |
| CN105662457A (en) * | 2016-03-22 | 2016-06-15 | 宁波元鼎电子科技有限公司 | Intelligent echometer |
| 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 |
| CN105943080A (en) * | 2016-06-30 | 2016-09-21 | 上海拓萧智能科技有限公司 | Intelligent stethophone |
| CN106308847A (en) * | 2016-08-31 | 2017-01-11 | 周辉 | Intelligent electronic stethoscope |
| CN107174277B (en) * | 2017-06-20 | 2020-08-04 | 深圳市泽科科技有限公司 | FPGA-based cardiopulmonary sound separation multi-path stethoscope |
| CN107945817B (en) * | 2017-11-15 | 2021-10-22 | 广东顺德西安交通大学研究院 | Method for classifying heart-lung sound signals, method for detecting heart-lung sound signals, device, medium and computer equipment |
| CN107693044A (en) * | 2017-11-15 | 2018-02-16 | 广东顺德西安交通大学研究院 | Surveillance of Coronary Heart diagnostic device |
| CN108323158A (en) * | 2018-01-18 | 2018-07-24 | 深圳前海达闼云端智能科技有限公司 | Heart sound identification method and cloud system |
| CN108784733B (en) * | 2018-06-28 | 2022-04-12 | 上海掌门科技有限公司 | Auscultation system and method |
| CN110123367B (en) * | 2019-04-04 | 2022-11-15 | 平安科技（深圳）有限公司 | Computer device, heart sound recognition method, model training device, and storage medium |
| CN110731778B (en) * | 2019-07-22 | 2022-04-29 | 华南师范大学 | Method and system for recognizing breathing sound signal based on visualization |
| US11463522B1 (en) * | 2019-07-31 | 2022-10-04 | United Services Automobile Association (Usaa) | System and method for efficiently transmitting signals with repeating structures |
| CN112017695A (en) * | 2020-03-04 | 2020-12-01 | 上海交通大学医学院附属上海儿童医学中心 | System and method for automatically identifying physiological sound |
| CN111938690A (en) * | 2020-08-07 | 2020-11-17 | 深圳市罗湖区妇幼保健院 | Intelligent heart sound auscultation system for children |
| CN112274174A (en) * | 2020-10-25 | 2021-01-29 | 贵州大学 | Intelligent electronic auscultation control system, method, storage medium and electronic stethoscope |
| CN112545547A (en) * | 2020-11-25 | 2021-03-26 | 北京积水潭医院 | Breath sound analysis method and breath sound analysis system |
| CN113448571B (en) * | 2021-07-13 | 2023-05-16 | 闽江学院 | MATLAB-based digital signal processor code rapid generation method |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN201492411U (en) * | 2009-08-26 | 2010-06-02 | 济南大学 | Heart sound diagnosis test device |
| CN102113894A (en) * | 2009-12-30 | 2011-07-06 | 湖南名鼎医疗器械有限公司 | Electronic timing and counting stethoscope |
| CN201939375U (en) * | 2011-01-28 | 2011-08-24 | 浙江好络维医疗技术有限公司 | Portable wireless signal display storage device for electrocardiogram, heart sounds and breath sounds |
| CN201939498U (en) * | 2011-01-28 | 2011-08-24 | 浙江好络维医疗技术有限公司 | Portable wireless electrocardio, cardiac sound and breath sound collecting terminal device |
| CN102365053A (en) * | 2009-04-03 | 2012-02-29 | 夏普株式会社 | Health monitoring method and system |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2000002486A1 (en) * | 1998-07-08 | 2000-01-20 | Cirrus Systems, Llc | Analytic stethoscope |
| WO2011141916A1 (en) * | 2010-05-13 | 2011-11-17 | Sensewiser Ltd. | Contactless non-invasive analyzer of breathing sounds |
-
2012
- 2012-05-08 CN CN201210140040.7A patent/CN102697520B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102365053A (en) * | 2009-04-03 | 2012-02-29 | 夏普株式会社 | Health monitoring method and system |
| CN201492411U (en) * | 2009-08-26 | 2010-06-02 | 济南大学 | Heart sound diagnosis test device |
| CN102113894A (en) * | 2009-12-30 | 2011-07-06 | 湖南名鼎医疗器械有限公司 | Electronic timing and counting stethoscope |
| CN201939375U (en) * | 2011-01-28 | 2011-08-24 | 浙江好络维医疗技术有限公司 | Portable wireless signal display storage device for electrocardiogram, heart sounds and breath sounds |
| CN201939498U (en) * | 2011-01-28 | 2011-08-24 | 浙江好络维医疗技术有限公司 | Portable wireless electrocardio, cardiac sound and breath sound collecting terminal device |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP3100675A1 (en) * | 2015-06-03 | 2016-12-07 | IMEDI PLUS Inc. | Method and system for recognizing physiological sound |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102697520A (en) | 2012-10-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102697520B (en) | Electronic stethoscope based on intelligent distinguishing function | |
| CN202801659U (en) | Electronic stethoscope based on intelligent identification function | |
| Leng et al. | The electronic stethoscope | |
| Thiyagaraja et al. | A novel heart-mobile interface for detection and classification of heart sounds | |
| Hadjileontiadis | Lung sounds: An advanced signal processing perspective | |
| US20210030390A1 (en) | Electronic stethoscope | |
| US20220007964A1 (en) | Apparatus and method for detection of breathing abnormalities | |
| CN108509823A (en) | The detection method and device of QRS complex | |
| Omarov et al. | Artificial Intelligence in Medicine: Real Time Electronic Stethoscope for Heart Diseases Detection. | |
| Xiang et al. | Research of heart sound classification using two-dimensional features | |
| Wang et al. | Acoustic-pressure sensor array system for cardiac-sound acquisition | |
| Zhao et al. | A CNN based human bowel sound segment recognition algorithm with reduced computation complexity for wearable healthcare system | |
| CN111789629A (en) | Breath sound intelligent diagnosis and treatment system and method based on deep learning | |
| CN201683910U (en) | Intelligent cardiopulmonary analyzing instrument | |
| Lee et al. | Restoration of lung sound signals using a hybrid wavelet-based approach | |
| Abeyratne et al. | Pitch-jitter analysis of snoring sounds for the diagnosis of sleep apnea | |
| CN113616188A (en) | Respiration monitoring method based on inaudible sound frequency modulation continuous wave | |
| Touahria et al. | Discrete Wavelet based Features for PCG Signal Classification using Hidden Markov Models. | |
| JP2015031889A (en) | Acoustic signal separation device, acoustic signal separation method, and acoustic signal separation program | |
| Kala et al. | An objective measure of signal quality for pediatric lung auscultations | |
| Karar et al. | Embedded heart sounds and murmurs generator based on discrete wavelet transform | |
| CN215349053U (en) | Congenital heart disease intelligent screening robot | |
| Mansurjonovich et al. | Creation of Algorithms for Constructions of Wavelet Models Suitable for Ecg-Signals | |
| CN112336369B (en) | Coronary heart disease risk index evaluation system of multichannel heart sound signals | |
| Arora et al. | A fusion framework based on cepstral domain features from phonocardiogram to predict heart health status |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20210419 Address after: No.5, row 4, chunheli, Yanhe Road, Lutai Town, Ninghe District, Tianjin Patentee after: Yang Baicheng Patentee after: Xia Ligeng Address before: Tianjin city Ninghe County Dabei 301500 Jian Gu Town outside the village on the west side of independence Patentee before: TIANJIN WOKANG TECHNOLOGY Co.,Ltd. |
|
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20141029 |