RU2574713C2 - Bronchial sound record analysis device - Google Patents

Abstract

FIELD: medicine.

SUBSTANCE: invention refers to medical equipment, namely bronchial sound record analysis devices. The device comprises a series breathing mask, microphone, acoustic amplifier, electronic filter box, analogue-to-digital converter and microcontroller connected to a computer though a communication hub, accessory microphone connected through an accessory analogue-to-digital converter to the microcontroller, acoustic signal filter, noise reduction unit, signal processing and analysis unit, decision support unit, local wireless communication hub connected to the computer, buffer storage keyboard, and buffer storage connected to the microcontroller; a colour touch-screen, external media and Internet- and LAN-connected local wireless communication hub connected to the computer are also provided. A digital colour video camera is connected to a separate input of the computer; a miniature electromechanical system configured attachable to the patient, comprising a sound-amplifying membrane, sound receiver and body, series-connected to the acoustic signal filter, noise reduction unit, signal processing and analysis unit, decision support unit, input and output connected to the microcontroller is connected to an input of the microcontroller. A series digital-to-analogue converter, telephone amplifier and head phones are connected to a separate output of the electromechanical system.

EFFECT: using the invention makes it possible to extend the range of facilities of the above application.

10 cl, 10 dwg

Description

The invention relates to a method for non-invasive (non-destructive, i.e., not requiring surgical intervention in the patient’s body) control of the state of the respiratory organs of a person or animal and to a device for its implementation. The scope covers medicine, where diagnostics of the state of human organs (bronchopulmonary diseases) is carried out, and can also be used in veterinary medicine to assess pathologies of animal organs.

Various non-invasive (non-destructive) devices for monitoring the condition of human bronchopulmonary organs are known and widely used. For example, single-channel or multi-channel electronic phonendoscopes with piezoelectric, optoelectronic or cardiomicrophone sensors [1-5].

An essential feature of these methods and devices is that the doctor makes a decision about the pathology by analyzing the acoustic signals displayed on the screen of the monitor, or by evaluating the characteristics of sound signals using head microphones.

The disadvantages of the methods and devices are large mass-dimensional characteristics (excluding the possibility of operation as a portable version), low noise immunity and low speed.

In addition, these methods and devices do not work automatically, because they only provide information about the state of the organ to the operator (doctor), according to which the specialist (narrow specialization doctor) must make a conclusion about the presence of pathologies of the human organ.

As a prototype, we consider an electronic bronchophonograph, which is designed as a prefix to a desktop computer (laptop) and contains a breathing mask with a miniature microphone, an acoustic amplifier, an electronic filter unit, an analog-to-digital converter, a microcontroller, a personal computer communication center, a set of processing programs, storage, search and visualization of the received data.

The disadvantages of the device are constructive and operational complexity, significant weight and size indicators, the need to use a personal computer to process the received data, search, store and visualize it, the need to conduct research in a specially prepared room, and the inability to operate as a wearable (mobile) version of the device. In addition, the method and device do not work in automatic mode.

The task is to create a device that improves the efficiency, speed, accuracy, functionality and convenience of diagnosing the anatomical and functional state of the respiratory organs of a person or animal.

The problem is achieved in that, as in the known and claimed device for electronic bronchophonography, receive and analyze acoustic signals, display and transmit information on the Internet.

The difference between the proposed device and the known one is that they form the reference and current matrices of acoustic signals about the pathology of organs, compare the reference and current matrices and automatically make a decision about the pathology during Wavelet, Fourier, fractal and correlation analysis of signals.

In addition, when analyzing the reference and current matrices, a complex comparison parameter is formed and analyzed.

In addition, the acoustic signals removed from the object are additionally filtered by a non-linear filter, suppress noise and approximate.

The task is also achieved by the fact that, like the known device, the claimed device for the analysis of bronchophonograms contains a breathing mask connected to the object in series, a microphone, an acoustic amplifier, an electronic filter unit, an analog-to-digital converter and a microcontroller connected via a node connected in series with bidirectional connections communications and a personal electronic computer to the computer, as well as a synchronizer and power supply.

The distinguishing feature of the proposed device from the known one is that the digital color video camera is connected to a separate computer input, and a miniature electromechanical system is connected to the microcontroller’s input. noise, signal processing and analysis unit, decision support unit, input and output connected to the microcontroller Ohm, a digital-to-analog converter, a telephone amplifier and headphones are connected in series to a separate output of the miniature electromechanical system.

In addition, the decision support unit comprises a series-connected interface, a memory unit, and a microprocessor.

In addition, the filter of acoustic signals is made in the form of a linear and adaptive filters.

In addition, the noise suppression unit is made in the form of a strip and non-linear signal converter.

In addition, the signal processing and analysis unit is made in the form of a spectral and time-frequency signal analyzer, non-linear and multi-speed signal converters.

In addition, the color touch screen is made in the form of a capacitive liquid crystal touch screen.

In addition, the external storage medium is made in the form of a flash drive.

In addition, the buffer drive is made in the form of a flash drive.

In addition, the keyboard of the buffer storage device is made in the form of a capacitive liquid crystal touch screen.

In addition, an additional universal power supply is made in the form of an integrated power source, combining a solar battery, a battery and an alternating current source.

The proposed device is illustrated in FIG. 1-10.

In FIG. 1 is a structural diagram of the device.

In FIG. 2 is a structural diagram of a miniature electromechanical system.

In FIG. 3 is a structural diagram of a decision support unit.

In FIG. 4 shows the mutual spectra of respiratory sounds on the surface of the chest of patients.

In FIG. 5 is a graph of 20 singular eigenvalues of a signal matrix.

In FIG. Figure 6 shows the original signal (in black) and the trend found using 10 singular approximation terms.

In FIG. 7 shows the results of phase-frequency tracking of the wave field of the signal.

In FIG. Figure 8 shows reference FAN images isolated for various classes of pathologies and for healthy breathing.

In FIG. Figure 9 shows an example of neural network training and testing.

In FIG. 10 shows graphs of generalization errors with confidence intervals in the test sample.

In FIG. 1 shows a structural diagram of a device that includes the following blocks:

1 - object (patient);

2 - respiratory mask;

3 - microphone;

4 - acoustic amplifier;

5 - block electronic filters (BEF);

6 - analog-to-digital Converter (ADC);

7 - digital color video camera (CCV);

8 - site local wireless communication (ULBS);

9 - additional external storage medium (DVD);

10 - an additional microphone;

11 - additional ADC;

12 - keyboard buffer buffer (KBN);

13 - buffer drive;

14 - built-in color touch screen (VCSE);

15 - computer;

16 - communication center with a personal computer (USPEVM);

17 - microcontroller;

18 - head phones;

19 - telephone amplifier;

20 - digital-to-analog converter;

21 - miniature electromechanical system (MEMS);

22 - decision support unit (BDPR);

23 - block processing and analysis of signals (BOAS);

24 - block noise reduction (BPSh);

25 - filter of acoustic signals;

26 - synchronizer;

27 - universal power supply.

In FIG. 1 shows the following relationships:

- электрические связи;

- electrical communications;

- многоканальные электрические связи;

- multichannel electrical communications;

- механические связи;

- mechanical bonds;

- акустические связи;

- acoustic communications;

- радиоканал.

- radio channel.

The device operates in two modes: preparation of a standard and control (diagnosis) of the patient's respiratory organs. The operating mode is set by the keyboard of the buffer storage device (KBN) 12 through the buffer storage device 13, the microcontroller 17, USPEVM 16 on the computer 15. In the preparation mode of the standard, a matrix (digital array) of acoustic signals is formed that characterize the state of a healthy organ. To do this, according to the signal from the synchronizer 26, all units of the device are turned on and the signal of object 1 passes through a breathing mask 2 and microphone 3, is fixed on object 1, to acoustic amplifier 4, which amplifies the electric signal. Next, this signal is filtered from interference by the electronic filter unit (BEF) 5 and through the microcontroller 17, the communication center with a personal computer (USPE) 16 is transmitted to the computer 15 for its subsequent analysis and assessment of the pathology of the patient's respiratory system. Simultaneously with the functioning of blocks 2-6, 11, 15, 16, a miniature electromechanical system (MEMS) 21 is activated, fixed on the object 1. MEMS 21 is placed in the focus of the sound lens of the sound receiver, provides additional signal amplification using a sound reinforcing membrane of the sound receiver (sound reinforcing membrane, sound receiver are part of MEMS 21). Since the membrane fits snugly to the patient’s body, and the sensor housing is made of special materials, the level of external noise is insignificant (see MEMS version, Fig. 2). The signals from MEMS 21 are received for analysis via the following three channels: firstly, the signal from MEMS 21 is digitized by a digital-to-analog converter (DAC) 20, amplified by a telephone amplifier 19 and displayed on the head phones 18; secondly, the signal from MEMS 21 enters through a microcontroller 17 and a communication center with a personal computer (USPE) 16 to a computer (microprocessor) 15, which according to a special program analyzes the signals describing the pathology of the patient's respiratory organs (object 1); thirdly, the signal from MEMS 21 is filtered from interference and noise by an acoustic signal filter (FAS) 25 and a noise suppression unit (BPS) 24, respectively, and is transmitted through the signal processing and analysis unit (BOAS) 23 to the decision support unit (microprocessor) ( BUPR) 22, the second input of which receives a signal from computer 15 through USPEIR 16 and microcontroller 17. BPR 22 generates incoming information from BOAS 23 and from computer 15 (through blocks 16, 17) as a reference data array (see BPR 22). Further, the reference information (from BPRP 22) is recorded through blocks 17, 16 into the memory of computer 15 or into an additional external storage medium (DVD) 9. Using the built-in color touch screen (WSCE) 14, the information in computer 15 is corrected and displayed on the screen of WSCE 14 A digital color video camera (CCV) 7 captures information (image) about the patient’s condition and other information (for example, recommendations for treatment from a doctor). This completes the operation of the device in the preparation mode of the standard (reference information - EI). In this case, the necessary information from the units of the device is transmitted through the wireless communication node (UPS) 8 to the Internet and the local computer network of the clinic.

In the control (diagnostic) mode, the computing units 2-26 of the device operate similarly to the standard preparation mode. In this mode, the current information (TI) corresponds to the already changed image of the respiratory organ (in the presence of pathology). This information is also recorded (like EI) in the memory of computer 15 and DVDR 9. Next, the EI and TI are compared and the degree of change in the patient’s organ is determined. This completes one cycle of the device.

In FIG. 2 shows a structural diagram of a miniature electromechanical system, which contains the following blocks:

28 - sound reinforcing membrane;

29 - a sound receiver;

30 - amplifier;

31 - case.

MEMS 21 operates as follows.

Structurally, MEMS 21 units are assembled in a housing 31, with the help of which a miniature electromechanical system is fixed to the patient's body (object 1). MEMS 21, placed in the focus of the sound lens of the sound receiver, provides additional signal amplification using a sound reinforcing membrane 28.

Since the membrane 28 is tightly attached to the patient’s body, and the sensor housing is made of special materials, the level of external noise is insignificant. Next, the sound signal from the ZOOM 28 is converted into an electrical signal using a sound receiver 29 and amplified by an amplifier 30.

FAS 25, BPSh 24 and BOAS 23 provide linear signal filtering (signal selection in the frequency domain), bandpass signal filtering (signal extraction in a given frequency band), adaptive signal filtering (signal recognition), spectral and time-frequency analysis, non-linear signal processing (correlation calculation, median filtering, vector coding), multi-speed signal processing (interpolation and decimation (reduction of the signal sampling frequency)) [7-10,17].

In FIG. 3 is a structural diagram of a decision support unit (BDP) 22, which consists of the following nodes:

32 - interface;

33 - memory block;

34 - microprocessor.

BUPR 22 (can be made in the form of a microprocessor AWP) works as follows.

Acoustic signals from the respiratory organs, i.e., from the respiratory organs, are received at interface 32 of the BDPR 22 from BOAS 23 and from the microcontroller from the respiratory mask and from MEMS 21. Thus, BPSR 22 forms two matrices in the following form:

- частотные характеристики сигнала (снимаемых с дыхательной маски 2 и МЭМС 21); F, F′ - Фурье-спектр мощности сигналов; G,G′ - амплитудные и геометрические характеристики огибающей сигнала (форма сигнала). При этом блок памяти 33 запоминает сначала эталонную информацию (в режиме подготовки эталона), а затем - текущую информацию (в режиме восприятия текущей информации).where the indices E and T mean the reference and current information, respectively;

- frequency characteristics of the signal (taken from the respiratory mask 2 and MEMS 21); F, F ′ - Fourier spectrum of signal power; G, G ′ are the amplitude and geometric characteristics of the envelope of the signal (waveform). In this case, the memory unit 33 first stores the reference information (in the preparation mode of the standard), and then the current information (in the perception mode of the current information).

Computer 15 is designed as an ARM microprocessor. The computer 15 in the pathology determination mode compares the reference and current information as follows:

where V is the sign of the comparison of matrices; I is a complex parameter for comparing the reference and current information by means of correlation, Fourier, wavelet, fractal and neural network analysis [7, 9, 10].

Let us consider the operation option of FAS 25, BPSh 24, BOAS 23, BPR 22 and computer 15. Let us give an example of isolating the spectral component of the structural analysis of respiration (sound) through tissues using low-pass filtering methods.

The separation of the spectral components of the air and structural conduction of respiratory sounds into the chest is shown in FIG. 4. Presented here:

a) above the lower parts of the lungs is normal,

b) above the upper parts of the lungs is normal,

c) above the lower parts of the lungs with pathology;

Re (W) is the real part of the mutual spectrum, Im (W) is the imaginary part of the mutual spectrum.

An analysis of the imaginary part of the mutual spectrum Im (W), characterizing the ratio of air and structural conduction [10], shows that the components of air conduction normally dominate in the frequency range from 100 to 300 Hz in the lower parts of the lungs (Fig. 2, a) and from 100 to 150-200 Hz in the upper ones (Fig. 2, b). Over areas with pathological deviations (Fig. 1, c), the replacement of air conduction components with structural conduction components in these spectral regions is observed, which may be a sign of local airway obstruction.

In FIG. 5 and FIG. Figure 6 shows examples of a singular matrix decomposition for extracting a reference signal.

In FIG. Figure 7 shows an example of constructing the amplitude wave field of the extracted reference signal, phase-frequency tracking of the wave field of the acoustic signal (the equilibrium processing algorithm of the amplitude signal).

In FIG. Figure 8 shows an example of the formation of reference FAN images of bronchopulmonary pathologies based on the calculation of quality functions (algorithm for nonequilibrium processing of the amplitude signal) (FAN - phase-time analysis).

It is presented here:

- class 0 - healthy breathing;

- class 1 - hard breathing;

- class 2 - sonorous wet (medium and fine bubble) rales;

- class 3 - crepitus;

- class 4 - wheezing;

- class 5 - dry rales;

- class 6 - trepid wheezing, intensifying during forced expiration.

In FIG. Figure 9 shows an example of neural network training and testing.

In FIG. 10 is a graph of a neural network learning error.

The graphs shown in FIG. 9-10, indicate the efficiency of the method and device using the example of the use of neural network learning technologies (recognition of respiratory pathologies)

The results presented in FIG. 4-10, obtained according to [9-17].

Compared with the known, the proposed method and device have a higher efficiency, speed and accuracy, as well as higher functionality and ease of diagnosis of the anatomical and functional state of the respiratory organs of a person or animal.

We show this by the example of the operation of the device.

The efficiency, speed and convenience of diagnosing the anatomical and functional state of human or animal organs is increased due to the implementation of a wearable (small-sized) version of the device and eliminating the need for diagnostics (research) in a specially prepared room. In addition, the convenience of the proposed device is provided by the introduction of a built-in color touch screen, an additional microphone, a digital color video camera, a buffer drive with a buffer drive keyboard, as well as the introduction of a communication function with the Internet network and the local area network of the clinic (due to a local wireless connection).

High accuracy characteristics of the proposed device are ensured, firstly, by removing additional acoustic information using a miniature electromechanical system, secondly, by introducing additional filtering, noise suppression, processing and analysis of acoustic signals, and thirdly, by a comprehensive (integral) comparison reference and current information due to correlation, Fourier, wavelet, fractal and neural network analysis (signals and information).

Information sources

1. Phonendoscope - an electronic stethoscope with a piezosensor. A.S. RU 2173538 C2 dated 04/05/1995.

2. Electronic phonendoscope with a cardiomicrophone. A.S. RU 2196512 C2 from 01/06/2000.

3. Multichannel electronic stethoscope. A.S. RU 2229843 C2 dated 12.24.2001.

4. Individual electronic stethoscope. Patent RU 2316256 C2 dated 01.22.2004.

5. Optoelectronic phonendoscope. Patent for utility model RU 86434 U1 dated 06/17/2009.

6. Electronic bronchophonograph. A.S. 2038041 from 09.16.1992 (Technical passport. Automated bronchophonographic diagnostic device "PATTERN - 01" according to TU9442 - 001 - 02066411 - 2008, No. РУ: 04789, registration certificate number: ФСР2009 / 04789, registration date 04/22/2009) (PROTOTYPE) .

7. Syryamkin V.I. and others. Modern methods of research of materials and nanotechnology. Textbook (Laboratory Workshop) / Ed. Doctor of Technical Sciences, Professor V.I. Syramykina. - Tomsk: Publishing house Tom. Univ., 2010 .-- 366 p.

8. Syryamkin V.I. and other Systems of technical vision: Handbook / Ed. IN AND. Syryamkina, B.C. Titotova. - Tomsk: MGP RASCO, 1992.367 s.

9. Syryamkin V.I. and other adaptive neural network algorithms for the diagnosis of materials, equipment and electronic equipment. - SaarbrUcken, Deutschland. Lambert - Academie Publishing, 2013 .-- 269 s.

10. Avedyuk O.A., Syryamkin V.I. and other Mathematical methods of computer science in problems and examples. Application experience in the design of complex systems. - Tomsk: Publishing house Tom. Univ., 2012 .-- 484 p.

11. Bauman E.V., Dorofeyuk A.A. Classification analysis of data // In: “Selected proceedings of the International Conference on Management Problems. Volume 1”. M .: SINTEG, 1999.

12. Bauman E.V., Zubchevsky N.S. The task of qualitative blurred classification // The Second International Conference on Management Problems. Abstracts of reports. Volume 1. M.: ILU RAS, 2003.S. 144.

13. Vovk I.V., Grinchenko V.T., Oleynik V.N. Problems of modeling the acoustic properties of the chest and measuring breathing noise // Acoust. journal 1995.Vol. 41, No. 5. S. 758-768.

14. Korenbaum V.I., Kulakov Yu.V., Tagiltsev A.A. The origin of the noise of vesicular respiration // Human Physiology. 1997. T. 23, No. 4. S. 133-135.

15. Kulakov Yu.V., Malyshenko I.Yu., Korenbaum V.I. Possibilities of combined bronchophonography in the diagnosis of pneumonia // Pulmonology. 2002.V. 12, No. 5. S. 29-32.

16. Kulakov Yu.V., Tagiltsev A.A., Korenbaum V.I., Kirichenko S.A. A device for studying the state of bronchial patency // Honey. Technics. 1995. No5. S. 20-23.

17. Tikhonov V.I. Statistical radio engineering // M .: Soviet radio, 1966. - 678 p.

Claims (10)

 1. A device for analyzing bronchophonograms containing a breathing mask, a microphone, an acoustic amplifier, an electronic filter unit, an analog-to-digital converter and a microcontroller connected to a computer through a communication node with it, an additional microphone connected through an additional analog-to-digital converter to the microcontroller, acoustic signal filter, noise suppression unit, signal processing and analysis unit, decision support unit, local wireless communication unit, connected included with a computer, a buffer drive keyboard and a buffer drive connected to the microcontroller, a color touch screen connected to the computer, an external storage medium and a local wireless communication node configured to access the Internet and a local area network, characterized in that the digital color the video camera is connected to a separate input of the computer, a miniature electromechanical system is connected to the input of the microcontroller, made with the possibility of fixing on the patient, containing a reinforcing membrane, a sound receiver and a housing and connected in series with an acoustic signal filter, a noise suppression unit, a signal processing and analysis unit, a decision support unit, an input and output connected to a microcontroller, connected in series to a miniature electromechanical system are a digital-to-analog converter, a telephone amplifier and headphones. 2. The device according to claim 1, characterized in that the decision support unit comprises a series-connected interface, a memory unit and a microprocessor. 3. The device according to claim 1, characterized in that the acoustic signal filter is made in the form of a linear and adaptive filters. 4. The device according to claim 1, characterized in that the noise suppression unit is made in the form of a strip and non-linear signal converter. 5. The device according to p. 1, characterized in that the signal processing and analysis unit is made in the form of a spectral and time-frequency signal analyzer, non-linear and multi-speed signal converters. 6. The device according to p. 1, characterized in that the color touch screen is made in the form of a capacitive liquid crystal touch screen. 7. The device according to p. 1, characterized in that the external medium is made in the form of a flash drive. 8. The device according to p. 1, characterized in that the buffer drive is made in the form of a flash drive. 9. The device according to p. 1, characterized in that the keyboard of the buffer drive is made in the form of a capacitive liquid crystal touch screen. 10. The device according to p. 1, characterized in that it further has a universal power supply unit, made in the form of an integrated power source that combines a solar battery, a battery and an alternating current source.

Images