Respiratory sound detection device, respiratory sound detection program, and respiratory sound detection system

Abstract

To provide a respiratory sound determination device for detecting abnormal respiration caused by an event such as a meal.SOLUTION: A respiratory sound determination device 20 determines a respiratory sound in an acoustic signal SS received from the outside. The respiratory sound determination device 20 includes a display device 26. The respiratory sound determination device 20 performs a first determination process when the start instruction is made, and performs a second determination process when the end instruction is made after the start instruction is made. The respiratory sound determination device 20 displays, on a display device 26, a first determination result indicating whether the respiratory sound in a first predetermined period is normal or abnormal in the first determination process, The respiratory sound determination device 20 displays, on the display device 26, a second determination result indicating whether the respiratory sound in a second predetermined period is normal or abnormal in the second determination process.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a breathing sound determination device that detects abnormal breathing caused by events such as meals.

Background Art

[0002] For example, Patent Document 1 discloses a breathing sound analysis system that detects abnormal breathing caused by aspiration or the like that occurs during sleep.

[0003] The breathing sound analysis system of Patent Document 1 includes a breathing sound analysis device (breathing sound determination device), a microphone, and a breathing band sensor. The microphone and the breathing band sensor are attached to a subject during sleep. Specifically, the microphone is attached to the throat of the subject, and the breathing band sensor is wound around the chest of the subject. The microphone converts the collected sound into an electrical signal (acoustic signal) and transmits it to the breathing sound determination device. The breathing band sensor converts the movement of the subject's chest into an electrical signal (operation signal) and transmits it to the breathing sound determination device. The breathing sound determination device continuously extracts the breathing sound of the subject based on the received acoustic signal and operation signal. The breathing sound determination device detects abnormal breathing by comparing a large number of extracted breathing sounds according to a predetermined algorithm. The detection result is displayed on the display device of the breathing sound determination device.

[0004] According to the breathing sound determination device of Patent Document 1, it is possible to detect abnormal breathing of a subject caused by events such as the intrusion of saliva into the throat. The caregiver of the subject can notice the abnormality of the subject's breathing through the display screen without staying with the subject during sleep.

Prior Art Documents

Patent Documents

[0005] <00​​​​​​​​​

[0006] Abnormal breathing can occur not only during sleep but also due to various events. For example, abnormal breathing can occur if foreign objects remain in the respiratory tract after eating. Furthermore, the frequency and depth of breathing can change due to factors such as exercise, bathing, or the use of medications like asthma inhalers, potentially causing or eliminating abnormal breathing. In other words, the presence or absence of abnormal breathing can change due to various events such as eating, exercise, bathing, and medication. Abnormal breathing may indicate a respiratory system abnormality. Especially in an aging society, there is a strong desire for easy detection of abnormal breathing caused by various events.

[0007] Therefore, the present invention aims to provide a respiratory sound detection device that detects abnormal breathing caused by events such as eating. [Means for solving the problem]

[0008] The present invention provides a first respiratory sound determination device, A respiratory sound detection device that determines respiratory sounds in an acoustic signal received from an external source, The aforementioned respiratory sound determination device is equipped with a display device, The respiratory sound detection device performs a first detection process when it is instructed to start, and performs a second detection process when it is instructed to stop after it has been instructed to start. In the first determination process, the respiratory sound determination device displays on the display device a first determination result indicating whether the respiratory sound during a first predetermined period is normal or abnormal. In the second determination process, the respiratory sound determination device displays on the display device the second determination result indicating whether the respiratory sound during the second predetermined period is normal or abnormal. We provide a respiratory sound detection device.

[0009] The present invention provides a second respiratory sound determination device, which is a first respiratory sound determination device, After performing the second determination process, the respiratory sound determination device transmits the first determination result and the second determination result to a different management device, depending on the combination of the first and second determination results. We provide a respiratory sound detection device.

[0010] The present invention provides a third respiratory sound determination device, which is a first respiratory sound determination device, The aforementioned respiratory sound detection device is a mobile terminal, The display device is the screen of the mobile terminal. We provide a respiratory sound detection device.

[0011] The present invention provides a fourth respiratory sound determination device, which is any of the first to third respiratory sound determination devices. If the respiratory sound detection device is unable to acquire the respiratory sound for a certain period of time after performing the first detection process but before being instructed to terminate, it will display information on the display device indicating that the respiratory sound could not be acquired. We provide a respiratory sound detection device.

[0012] The present invention provides a fifth respiratory sound determination device, which is any of the first to third respiratory sound determination devices. The respiratory sound determination device can perform the first determination process for the respiratory sounds in two or more acoustic signals whose first predetermined periods overlap, and can perform the second determination process for the respiratory sounds in two or more acoustic signals whose second predetermined periods overlap. We provide a respiratory sound detection device.

[0013] The present invention is a program for causing a mobile terminal to function as one of the first to third respiratory sound detection devices, as a first respiratory sound detection program. We provide a program for analyzing respiratory sounds.

[0014] The present invention provides a respiratory sound determination system comprising a respiratory sound determination device (one of the first to third) and two or more acoustic sensors, as a first respiratory sound determination system. Each of the acoustic sensors collects ambient sound when it starts operating, converts it into the acoustic signal, and transmits it externally. The breathing sound determination device receives the acoustic signals transmitted by two or more of the acoustic sensors. A breathing sound determination system is provided.

[0015] As a first monitoring method, the present invention provides a monitoring method for monitoring two or more subjects using a first breathing sound determination system, where the acoustic sensors are respectively attached to the subjects, before all of the subjects perform a predetermined event, the acoustic sensors are started, and a start instruction is given to the breathing sound determination device, after all of the subjects perform a predetermined event, an end instruction is given to the breathing sound determination device. A monitoring method is provided.

[0016] As a first acoustic sensor, an acoustic sensor having a sensor function of collecting sound and converting it into an acoustic signal which is an electrical signal, and a transmission function of transmitting the acoustic signal, the acoustic sensor includes a sensor body and a base member, the sensor body has a contact portion, the contact portion protrudes downward from the lower surface of the sensor body in the vertical direction, and has an annular shape in a horizontal plane orthogonal to the vertical direction, a recess is formed in the sensor body, the recess is recessed upward from the lower surface of the sensor body, and is surrounded by the contact portion in the horizontal plane, a sound collecting sheet is provided at the upper part of the recess, a through hole is formed in the base member, the sensor body is attached to the base member such that the contact portion passes through the through hole. An acoustic sensor is provided.

Advantages of the Invention

[0017] The respiratory sound detection device of the present invention can display two detection results: a first detection result for a first predetermined period, and a second detection result for a second predetermined period that follows the first predetermined period. Therefore, by issuing a start command before an event such as a meal and an end command after the event, the detection results before and after the event can be compared. More specifically, if the detection result before a meal is normal and the detection result after a meal is abnormal, it can be determined that abnormal breathing occurred due to the meal. For example, it can be determined that there may be foreign matter remaining in the upper part of the subject's respiratory organ due to the meal. In other words, the present invention provides a respiratory sound detection device that can detect abnormal breathing caused by events such as meals. [Brief explanation of the drawing]

[0018] [Figure 1] This figure shows a respiratory sound determination system according to an embodiment of the present invention. [Figure 2] Figure 1 is a perspective view showing the acoustic sensor of the respiratory sound detection system. The position of the lamp housed inside the sensor body is indicated by a dashed line. [Figure 3] Figure 2 is an exploded perspective view showing the acoustic sensor. The position of the lamp is indicated by a dashed line. [Figure 4] Figure 3 is a bottom view showing the sensor body of the acoustic sensor. [Figure 5] Figure 3 is a front view showing the sensor body of the acoustic sensor. A portion of the base material's contour is drawn with a dashed line. A portion of the acoustic sensor (enclosed by a dashed line) is shown in a magnified view. In the magnified view, the contour of the hidden recess is drawn with a dashed line. [Figure 6] Figure 1 is a block diagram showing the respiratory sound detection device of the respiratory sound detection system. [Figure 7] Figure 6 shows the structure of acoustic data stored in the auxiliary memory of the respiratory sound detection device. [Figure 8] Figure 6 shows the timing at which the respiratory sound detection device detects respiratory sounds. [Figure 9]Figure 6 is a flowchart showing the control process of the respiratory sound detection program for the respiratory sound detection device. [Figure 10] Figure 9 is a flowchart showing the acoustic signal storage process of the respiratory sound detection program. [Figure 11] Figure 9 is a flowchart showing the first determination process of the respiratory sound determination program. [Figure 12] Figure 9 is a flowchart showing the second determination process of the respiratory sound determination program. [Figure 13] Figures 11 and 12 illustrate the input / output layers of the anomaly detection model. [Figure 14] This figure shows an example of a convolutional network in the anomaly detection model shown in Figure 13. [Figure 15] Figures 11 and 12 show an example of a method for generating an anomaly detection model. [Figure 16] This figure shows a modified version of the respiratory sound detection system shown in Figure 1. [Figure 17] Figure 16 shows the timing at which the respiratory sound detection system detects respiratory sounds. [Modes for carrying out the invention]

[0019] Referring to Figure 1, the respiratory sound determination system 10 according to an embodiment of the present invention is a system that determines whether the respiratory sounds of a subject 70 are normal or abnormal. The respiratory sound determination system 10 of this embodiment is used in a nursing care facility. The subject 70 in this embodiment is an elderly person residing in a nursing care facility. However, the present invention is not limited to this and can be applied in various ways. For example, the respiratory sound determination system 10 may be used in a sports gym or in an ordinary home. For example, the respiratory sound determination system 10 may be used for home nursing care in an ordinary home.

[0020] The respiratory sound determination system 10 of this embodiment includes a respiratory sound determination device 20, a management device 40, and an acoustic sensor 60.

[0021] The acoustic sensor 60 is a small electronic device that can be attached to and removed from the subject 70. More specifically, the acoustic sensor 60 can be attached to the skin of the subject 70's throat. The acoustic sensor 60 is equipped with various electronic components, such as a microphone (not shown). The microphone of the acoustic sensor 60 collects sounds such as breathing sounds generated in the subject 70's airway and converts them into an acoustic signal SS, which is an electrical signal. The acoustic sensor 60 of this embodiment transmits the acoustic signal SS via short-range wireless communication such as Bluetooth®.

[0022] As described above, the acoustic sensor 60 of this embodiment has a sensor function that collects sound and converts it into an acoustic signal SS, and a transmission function that transmits the acoustic signal SS wirelessly over short range. However, the present invention is not limited thereto. For example, the acoustic sensor 60 may have other functions in addition to the sensor function and the transmission function. The acoustic sensor 60 may transmit the acoustic signal SS wirelessly using a method other than wireless communication over short range, or it may transmit it wirelessly via a communication cable.

[0023] In this embodiment, the respiratory sound detection device 20 and the management device 40 are mobile terminals such as smartphones. Both the respiratory sound detection device 20 and the management device 40 are easily portable. However, the present invention is not limited thereto. For example, at least one of the respiratory sound detection device 20 and the management device 40 may be a portable PC (personal computer) or a desktop PC.

[0024] The respiratory sound detection device 20 is capable of receiving the acoustic signal SS. More specifically, the respiratory sound detection device 20 in this embodiment is paired with the acoustic sensor 60 and is positioned at a distance from which it can receive the acoustic signal SS. However, the present invention is not limited thereto. For example, if the acoustic signal SS is transmitted via a wired connection through a communication cable, the respiratory sound detection device 20 may be connected to the communication cable.

[0025] The respiratory sound detection device 20 can transmit various data to the control device 40. The control device 40 can receive the data transmitted by the respiratory sound detection device 20. The method of transmission and reception between the respiratory sound detection device 20 and the control device 40 is not particularly limited.

[0026] The respiratory sound determination device 20 determines the respiratory sounds in the acoustic signal SS received from an external source. Specifically, the respiratory sound determination device 20 performs two determination processes: a first determination process and a second determination process that follows the first determination process. In both the first and second determination processes, the respiratory sound determination device 20 determines and displays whether the respiratory sounds of the subject 70 contained in the received acoustic signal SS are normal or abnormal. The respiratory sound determination device 20 also transmits the determination results of the first and second determination processes to the management device 40.

[0027] In this embodiment, the respiratory sound detection device 20 is operated by a caregiver at the nursing facility, and the management device 40 is operated by the administrator of the nursing facility. For example, in the first detection process, the respiratory sound detection device 20 detects the respiratory sound of the subject 70 before a meal, and in the second detection process, it detects the respiratory sound of the subject 70 after a meal. For example, if the respiratory sound before a meal is normal and the respiratory sound after a meal is abnormal, the caregiver and administrator can know that the abnormal breathing was caused by the meal, and thereby provide appropriate care to the subject 70. For example, the caregiver can remove foreign objects from the subject 70's respiratory system by having the subject 70 gargle. The administrator can also consider whether it is necessary to change the contents of the meal.

[0028] As described above, the operators of the respiratory sound detection device 20 and the management device 40 in this embodiment are caregivers and administrators of a nursing care facility, respectively. However, the present invention is not limited thereto, and the operators of the respiratory sound detection device 20 and the management device 40 may differ depending on the application of the respiratory sound detection device 20. For example, when the respiratory sound detection system 10 is used in a typical home, a family member of the subject 70 may operate the respiratory sound detection device 20, or a staff member of a remote nursing care facility may operate the management device 40.

[0029] The respiratory sound detection system 10 shown in Figure 1 comprises one respiratory sound detection device 20, one management device 40, and one acoustic sensor 60. However, the present invention is not limited thereto. For example, the respiratory sound detection system 10 may further include other devices in addition to the above-described devices. The respiratory sound detection system 10 may also include two or more respiratory sound detection devices 20, and two or more acoustic sensors 60. Furthermore, the respiratory sound detection system 10 may not include a management device 40.

[0030] The acoustic sensor 60 of this embodiment will now be described.

[0031] Referring to Figures 2 and 3, the acoustic sensor 60 of this embodiment comprises a sensor body 62 and a base member 67. The sensor body 62 is the component responsible for the sensor function and transmission function of the acoustic sensor 60. The base member 67 is attached to the skin of the throat of the subject 70 (see Figure 1).

[0032] Referring to Figures 3 to 5, the sensor body 62 of this embodiment has a thin rectangular flat plate shape. Specifically, the size of the sensor body 62 in the front-to-back direction (X direction) is approximately 45 mm, and the size of the sensor body 62 in the lateral direction (Y direction) perpendicular to the front-to-back direction is approximately 35 mm. Furthermore, the size of the sensor body 62 in the up-and-down direction (Z direction) perpendicular to both the front-to-back and lateral directions is approximately 8.5 mm. Terms such as up-and-down direction do not indicate an absolute positional relationship with respect to the ground, but merely indicate the relative positional relationship when the part of the surface of the sensor body 62 that faces the skin of the subject 70 (see Figure 1) is defined as the bottom surface.

[0033] The sensor body 62 of this embodiment includes a switch 63, a lamp 64, a connector 65, and a contact portion 66, in addition to various electronic components such as a microphone (not shown) housed inside the sensor body 62. Furthermore, four locking portions 624 are formed on the upper surface (+Z side) of the sensor body 62, and a recess 662 is formed on the lower surface (-Z side) of the sensor body 62.

[0034] Referring to Figure 3, switch 63 is the power switch for the sensor body 62. Pressing switch 63 turns it ON, and pressing it again turns it OFF. The sensor body 62 operates when switch 63 is ON. Lamp 64 is housed inside the sensor body 62 and lights up when the sensor body 62 is operating. When lit, lamp 64 emits light upwards from the sensor body 62 through a translucent material. Connector 65 is used to charge the sensor body 62.

[0035] Referring to Figure 2 in conjunction with Figure 1, the switch 63 and lamp 64 in this embodiment are also used when pairing the respiratory sound detection device 20 with the acoustic sensor 60. More specifically, when the switch 63 is pressed and held for about 5 to 7 seconds, the acoustic sensor 60 enters a ready state, and the lamp 64 flashes blue and red. When pairing is complete, the acoustic sensor 60 becomes ready for use, and the lamp 64 lights up red. However, the present invention is not limited thereto. For example, the conditions under which the lamp 64 lights up or flashes, and the color of the lamp 64, can be modified as needed.

[0036] Referring to Figures 4 and 5, the contact portion 66 protrudes downward from the lower surface of the sensor body 62 in the vertical direction and has an annular shape in a horizontal plane (XY plane) perpendicular to the vertical direction. In this embodiment, the contact portion 66 has a continuous annular shape in the horizontal plane. However, the present invention is not limited thereto. For example, the contact portion 66 may have a continuous annular shape or a rectangular annular shape in the horizontal plane.

[0037] The recess 662 is recessed upward from the lower surface of the sensor body 62 and is surrounded by the contact portion 66 in the horizontal plane. A sound-collecting sheet 664 is provided on the upper part of the recess 662. The sound-collecting sheet 664 vibrates when it receives sound. The sound-collecting sheet 664 in this embodiment has a circular shape in the horizontal plane and vibrates uniformly. However, the present invention is not limited thereto. For example, the shape of the sound-collecting sheet 664 can be changed as needed.

[0038] The sensor body 62 of this embodiment has the structure described above. However, the present invention is not limited thereto, and the structure of the sensor body 62 can be modified as needed.

[0039] Referring to Figure 3, the base member 67 has a thin, flexible bottom portion 672 and four locking portions 674. The bottom portion 672 extends along the horizontal plane and has a through hole 68 formed therein. That is, the base member 67 has a through hole 68. The through hole 68 has a circular shape in the horizontal plane that corresponds to the contact portion 66 (see Figure 4) and penetrates the bottom portion 672 in the vertical direction. The four locking portions 674 surround the through hole 68 in the horizontal plane and protrude upward from the bottom portion 672. A tag 69 is also attached to the base member 67. For example, the tag 69 has the identification name of the acoustic sensor 60 and a two-dimensional code corresponding to the identification name printed on it.

[0040] The base member 67 of this embodiment has the structure described above. However, the present invention is not limited thereto, and the structure of the base member 67 can be modified as needed. For example, the tag 69 may be provided as needed.

[0041] Referring to Figures 2 and 3 in conjunction with Figure 5, the sensor body 62 is attached to the base member 67 such that the contact portion 66 passes through the through hole 68. When the sensor body 62 is attached to the base member 67, the locking portion 674 locks the locked portion 624. At this time, the lower surface of the contact portion 66 is substantially flush with the lower surface of the bottom portion 672. When the bottom portion 672 is attached to the skin of the subject 70 (see Figure 1), the sound collection sheet 664 faces the skin of the subject 70.

[0042] According to this embodiment, the base member 67 is attached to the skin of the throat of the subject 70 (see Figure 1), and the sound-collecting sheet 664 faces the skin of the subject 70's throat. The sound-collecting sheet 664, positioned in this manner, can reliably transmit the respiratory sounds generated in the subject 70's airway to the inside of the sensor body 62. On the other hand, background noise such as conversations occurring around the subject 70 is less likely to be transmitted to the inside of the sensor body 62. Therefore, the acoustic sensor 60 can effectively collect the respiratory sounds generated in the subject 70's throat while suppressing background noise.

[0043] Referring to Figure 2, the acoustic sensor 60 of this embodiment includes a noise cancellation circuit (not shown) and an external microphone 622. The noise cancellation circuit is mounted inside the sensor body 62. The external microphone 622 has an opening on the top surface of the sensor body 62 and is connected to the noise cancellation circuit. The external microphone 622 converts noise such as conversation occurring on the opposite side of the subject's throat into a signal and transmits it to the noise cancellation circuit. The noise cancellation circuit generates a signal with the opposite phase to the signal transmitted from the external microphone 622 to cancel the noise. As described above, the acoustic sensor 60 of this embodiment has a noise cancellation function and can selectively collect breathing sounds occurring in the throat of the subject 70.

[0044] The acoustic sensor 60 of this embodiment has the structure described above and is attached to the subject 70 (see Figure 1) as described above. However, the present invention is not limited thereto. For example, the acoustic sensor 60 may be attached to a part of the subject 70 other than the skin of the throat.

[0045] The respiratory sound detection device 20 of this embodiment (see Figure 1) will be described below.

[0046] Referring to Figure 6, the respiratory sound detection device 20 of this embodiment comprises a processing device 21, a main memory 22, an auxiliary memory 23, an input device 25, a display device 26, an antenna 28, and an additional antenna 29. The processing device 21 is communicatively connected to other devices via an internal bus. The illustrated respiratory sound detection device 20 comprises only the above-described devices. However, the present invention is not limited thereto. For example, the respiratory sound detection device 20 may further include other devices such as a speaker in addition to the above-described devices.

[0047] According to this embodiment, each of the above-described devices is part of a single mobile terminal. For example, the display device 26 is the screen of the mobile terminal. The input device 25 is a virtual input unit, such as a keyboard or input buttons, displayed on the screen of the mobile terminal. The antenna 28 is a short-range communication antenna mounted inside the mobile terminal. The antenna 28 receives the acoustic signal SS transmitted by the acoustic sensor 60 paired with the respiratory sound determination device 20. The additional antenna 29 is a communication antenna mounted inside the mobile terminal. The additional antenna 29 transmits the respiratory sound determination result of the subject 70 (see Figure 1) to the management device 40.

[0048] The respiratory sound detection device 20 of this embodiment has the structure described above. However, the present invention is not limited thereto. For example, if the respiratory sound detection device 20 is a PC, the display device 26 may be a liquid crystal display separate from the processing device 21, and the input device 25 may be a keyboard or mouse separate from the processing device 21. In this case, the input device 25 and the display device 26 may each be connected to the processing device 21 via a communication cable so as to be able to communicate.

[0049] The processing unit 21 is equipped with a CPU (Central Processing Unit: not shown). The auxiliary storage device 23 is, for example, an EEPROM (Electrically Erasable Programmable Read-Only Memory) and can store various files, including the executable file of the respiratory sound detection program 30 and acoustic data 32. The auxiliary storage device 23 retrieves and stores files in response to instructions from the processing unit 21. The CPU of the processing unit 21 retrieves the executable file stored in the auxiliary storage device 23, loads it into the main memory 22, and executes the instructions within the executable file to realize various functions.

[0050] The input device 25 transmits the input characters and the indicated position or range to the processing device 21. The display device 26 displays the characters and images transmitted from the processing device 21. For example, the CPU (not shown) of the processing device 21 loads the respiratory sound detection program 30 into the main memory 22 in response to the instructions input from the input device 25 and executes the respiratory sound detection process. That is, the respiratory sound detection program 30 is a program that enables the mobile terminal to function as a respiratory sound detection device 20. The respiratory sound detection program 30 is installed in the auxiliary memory 23, for example, via the WEB (World Wide Web).

[0051] As described above, the CPU (not shown) of the processing unit 21 actually executes programs such as the respiratory sound detection program 30. However, in the following explanation, it may be described as if the respiratory sound detection program 30 is the main execution entity, or as if the CPU processing of the respiratory sound detection program 30 is the main execution entity.

[0052] Referring to Figure 7 in conjunction with Figure 6, the acoustic data 32 is a file that stores the acoustic signal SS received by the respiratory sound detection device 20. In this embodiment, the acoustic data 32 is created when the respiratory sound detection program 30 starts operating and deleted before the respiratory sound detection program 30 finishes operating. For this reason, the file name of the acoustic data 32 may be a fixed name. However, the present invention is not limited thereto. For example, the respiratory sound detection program 30 does not have to delete the acoustic data 32. That is, the acoustic data 32 may be added each time the respiratory sound detection program 30 starts operating. In this case, the file name of the acoustic data 32 may be a fixed name and the date of creation.

[0053] As described later, the acoustic data 32 is empty when it is created. The acoustic data 32 stores a start record 32S and an end record 32E as the respiratory sound determination program 30 processes. The start record 32S and the end record 32E each store partial acoustic data 328, which is a part of the acoustic signal SS received by the respiratory sound determination device 20, the start time 322 of the partial acoustic data 328, and the end time 324 of the partial acoustic data 328. The acoustic data 32 in this embodiment has the above-described data structure. However, the present invention is not limited thereto. The data structure of the acoustic data 32 can be modified as needed. Also, the acoustic data 32 may be provided as needed.

[0054] The respiratory sound detection program 30 of this embodiment will be described below.

[0055] The respiratory sound determination program 30 of this embodiment performs four processes: control processing (see Figure 9), acoustic signal storage processing (see Figure 10), first determination processing (see Figure 11), and second determination processing (see Figure 12). For example, the control processing, first determination processing, and second determination processing are executed on one thread, while the acoustic signal storage processing is executed on a separate thread. Once the acoustic signal storage processing starts, it continues to store the acoustic signal SS received by the antenna 28 in the main memory 22 until the start of the second determination processing. However, the present invention is not limited thereto. The structure and algorithm of the respiratory sound determination program 30 can be modified as needed.

[0056] The control process of this embodiment will be described below.

[0057] Referring to Figure 9 in conjunction with Figures 1, 6, and 7, once the respiratory sound detection program 30 starts operating, the control process continues to wait until the operator gives a start command (S910). For example, the control process displays an input button (start button) with "Start Recording" on the display device 26 and continues to wait for the start button to be pressed.

[0058] If the control process is instructed to start by pressing the start button (YES in S910), it creates new empty acoustic data 32 (i.e., acoustic data 32 that does not include the start record 32S and the end record 32E) (S915). Next, the control process starts the acoustic signal storage process (S920). As will be described later, once the acoustic signal storage process starts, it stores the start record 32S in the acoustic data 32 based on the received acoustic signal SS.

[0059] Next, the control process waits until the acoustic signal SS of CP is stored in the acoustic data 32 for a certain period of time (S930). The certain period of time CP is, for example, 30 seconds. More specifically, the control process waits until the start record 32S is recorded by the acoustic signal storage process. Once the start record 32S is stored, the control process performs the first determination process (S940). As will be described later, the first determination process provides a first determination result indicating whether the respiratory sound stored in the start record 32S is normal or abnormal. The first determination result is displayed on the display device 26.

[0060] Referring to Figure 8 in conjunction with Figure 1, according to this embodiment, the operator of the respiratory sound detection device 20 gives the start command before the subject 70 eats. During a certain period of CP, there is virtually no food remaining in the subject 70's mouth, and therefore, no abnormal breathing caused by eating occurs.

[0061] Referring to Figure 9 in conjunction with Figures 1, 6, and 7, the control process then continues to wait until the operator gives a termination command (S950). For example, the control process displays an input button (termination button) with "Start Analysis" on the display device 26 and continues to wait for the termination button to be pressed.

[0062] If the control process is terminated by pressing the terminate button (YES in S950), it notifies the acoustic signal storage process that it has been terminated (S960). As will be described later, when the acoustic signal storage process is terminated, it stores a termination record 32E in the acoustic data 32 based on the received acoustic signal SS.

[0063] Next, the control process performs a second determination process (S970). As will be described later, the second determination process provides a second determination result indicating whether the respiratory sound stored in the termination record 32E is normal or abnormal. The second determination result is displayed on the display device 26.

[0064] Referring to Figure 8 in conjunction with Figure 1, according to this embodiment, the operator of the respiratory sound detection device 20 gives a termination command after the subject 70 has finished eating. The operator may remove the acoustic sensor 60 from the subject 70 before giving the termination command, or may remove the acoustic sensor 60 from the subject 70 after giving the termination command.

[0065] Referring to Figure 9 in conjunction with Figures 1, 6, and 7, the control process then transmits the determination result to the management device 40 (S980). In this embodiment, the control process transmits both the first and second determination results depending on the combination of the first and second determination results. For example, the control process transmits both the first and second determination results only if the first determination result is "normal" and the second determination result is "abnormal". In this case, the operator of the management device 40 can know that abnormal breathing has occurred due to eating, and can then provide the necessary care to the subject 70.

[0066] Next, the control process deletes the acoustic data 32 (S990) and terminates its operation.

[0067] The acoustic signal storage process of this embodiment will be described below.

[0068] Referring to Figure 10 in conjunction with Figures 1, 6, and 7, the acoustic signal storage process, once it starts, continues to receive the acoustic signal SS into the main memory 22 while determining whether the acoustic signal SS has been interrupted (S1010) until it receives the acoustic signal SS of CP for a certain period of time (S1015). If the acoustic signal SS has not been interrupted, i.e., if the acoustic signal SS can be received (NO in S1010), the acoustic signal storage process determines whether the acoustic signal SS of CP has been received for a certain period of time (S1015). If the acoustic signal storage process has not been received for a certain period of time (NO in S1015), it continues to receive the acoustic signal SS (S1010).

[0069] If the acoustic signal storage process has received the acoustic signal SS of CP for a certain period of time (if the answer is YES in S1015), it determines whether or not the received acoustic signal SS contains breathing sounds (S1020). If the acoustic signal storage process finds that breathing sounds are present (if the answer is YES in S1020), it records the received acoustic signal SS of CP for a certain period of time in the start record 32S of the acoustic data 32 (S1025).

[0070] More specifically, the acoustic signal storage process adds a start record 32S to the acoustic data 32. At this time, the acoustic signal storage process stores the start time 322, end time 324, and partial acoustic data 328 of the start record 32S. The acoustic signal storage process stores the acoustic signal SS of CP in the partial acoustic data 328 for a certain period of time. The acoustic signal storage process stores the time when reception of the acoustic signal SS to be stored began as the start time 322. The acoustic signal storage process stores the time when the acoustic signal SS to be stored was last received as the end time 324.

[0071] The acoustic signal storage process of this embodiment determines that a breath sound is present in an acoustic signal SS if its intensity is high for a certain period of time. More specifically, the acoustic signal storage process determines that a breath sound is present if the amplitude of the received acoustic signal SS exceeds a predetermined value, even partially. However, the present invention is not limited thereto. For example, the acoustic signal storage process may use a neural network model to determine whether or not a breath sound is present.

[0072] The acoustic signal storage process stores the acoustic signal SS in the acoustic data 32 and continues to receive the acoustic signal SS into the main memory 22 while determining whether the acoustic signal SS has been interrupted (S1050) until it is notified by the control process that it has been instructed to terminate (S1060). If the acoustic signal SS has not been interrupted (NO in S1050), the acoustic signal storage process determines at predetermined unit time intervals (for example, 2 seconds) whether the acoustic signal SS of the last received period CP contains a breathing sound (S1055). If a breathing sound is present (YES in S1055), the acoustic signal storage process determines whether it has been instructed to terminate by the control process (S1060). If it has not been instructed to terminate (NO in S1060), the acoustic signal storage process continues to receive the acoustic signal SS (S1050).

[0073] If the control process instructs it to terminate (YES in S1060), the acoustic signal storage process records the last received acoustic signal SS for a certain period of time CP in the termination record 32E of the acoustic data 32 (S1065), and then terminates the process.

[0074] If the acoustic signal SS is interrupted (NO in S1010 or S1050), the acoustic signal storage process displays a confirmation message on the display device 26 (S1030 or S1070). For example, the acoustic signal storage process displays a message such as "No sound is being received from the sensor. Please check the sensor," along with a continue button and an end button on the display device 26, which are input buttons for selecting whether to continue or end the process.

[0075] If the received acoustic signal SS does not contain breathing sounds (NO in S1020 or S1055), the acoustic signal storage process displays a confirmation message on the display device 26 (S1040 or S1080). For example, the acoustic signal storage process displays a message such as "No breathing sounds were detected. Please check the sensor," along with a continue button and an end button on the display device 26, which are input buttons for selecting whether to continue or end the process.

[0076] The operator of the respiratory sound detection device 20 can learn from the confirmation message described above that a problem has occurred with the acoustic sensor 60. For example, if the switch 63 (see Figure 2) of the acoustic sensor 60 is in the OFF position, the operator can press the switch 63 to turn it ON. If the acoustic sensor 60 has come off, the operator can reattach it. The operator can also press the continue button or the end button depending on the situation.

[0077] If the continue button is pressed (YES in S1035, S1045, S1075, or S1085), the acoustic signal storage process continues to receive the acoustic signal SS again (S1010 or S1050). On the other hand, if the end button is pressed (NO in S1035, S1045, S1075, or S1085), the acoustic signal storage process notifies the control process of an error termination and terminates the process ("A"). Referring to Figure 9 in conjunction with Figure 6, when the control process is notified of an error termination ("A"), it deletes the acoustic data 32 (S990) and terminates the process.

[0078] Referring to Figures 6 and 7, as can be understood from the above explanation, the start record 32S of the acoustic data 32 stores the acoustic signals SS received at CP for a certain period after the start instruction. The end record 32E of the acoustic data 32 stores the acoustic signals SS received at CP for a certain period before the end instruction. According to this embodiment, the necessary acoustic signals SS can be stored in the acoustic data 32 by a relatively simple acoustic signal storage process. For this reason, the respiratory sound determination program 30 can easily determine respiratory sounds.

[0079] However, the present invention is not limited to this embodiment, and the acoustic signal storage process can be modified as needed. For example, the acoustic signal storage process may store all acoustic signals SS received between the start instruction and the end instruction in acoustic data 32 at regular intervals CP.

[0080] The first and second determination processes of this embodiment will be described below.

[0081] Referring to Figure 11 in conjunction with Figures 6 and 7, the first determination process first creates the input layer of the anomaly discrimination model (S1110). As will be described later, the anomaly discrimination model is a neural network generated through machine learning. More specifically, the first determination process creates the input layer of the anomaly discrimination model based on the partial acoustic data 328 of the start record 32S stored in the acoustic data 32 (i.e., the acoustic signal SS of CP for a certain period of time). Next, the first determination process obtains the output layer of the anomaly discrimination model using the created input layer (S1120). Next, the first determination process creates the first determination result from the output layer of the anomaly discrimination model (S1130). The first determination result indicates whether the breathing sound contained in the acoustic signal SS of CP for a certain period of time is normal or abnormal. Next, the first determination process displays the first determination result on the display device 26 (S1140).

[0082] Referring to Figure 1, the first judgment result in this embodiment is displayed as "normal" or "abnormal". If "abnormal" is displayed, the operator of the respiratory sound detection device 20 can provide necessary care to the subject 70. However, the present invention is not limited thereto. For example, the first judgment result may be displayed as "○" or "×".

[0083] Referring to Figure 12 in conjunction with Figures 6 and 7, the second determination process first creates the input layer of the anomaly detection model (S1210). More specifically, the second determination process creates the input layer of the anomaly detection model based on the partial acoustic data 328 of the termination record 32E stored in the acoustic data 32 (i.e., the acoustic signal SS of CP for a certain period of time). Next, the second determination process obtains the output layer of the anomaly detection model using the created input layer (S1220). Next, the second determination process creates the second determination result from the output layer of the anomaly detection model (S1230). The second determination result is CP for a certain period This result indicates whether the respiratory sounds contained in the acoustic signal SS are normal or abnormal. Next, the second determination process displays the second determination result on the display device 26 (S1240).

[0084] Referring to Figure 1, the second judgment result in this embodiment is displayed as "normal" or "abnormal". If "abnormal" is displayed, the operator of the respiratory sound detection device 20 can provide necessary care to the subject 70. However, the present invention is not limited thereto. For example, the second judgment result may be displayed as "○" or "×".

[0085] Referring to Figures 1, 6, and 8, the above description of the respiratory sound determination device 20 can be summarized as follows: When a start command is given, the respiratory sound determination device 20 of this embodiment performs a first determination process, and when a stop command is given after a start command, it performs a second determination process. In the first determination process, the respiratory sound determination device 20 displays a first determination result on the display device 26 indicating whether the respiratory sound during a first predetermined period is normal or abnormal. In the second determination process, the respiratory sound determination device 20 displays a second determination result on the display device 26 indicating whether the respiratory sound during a second predetermined period is normal or abnormal.

[0086] In this embodiment, the first predetermined period and the second predetermined period are each equal to the constant period CP in the acoustic signal storage processing (see Figure 10). That is, the first predetermined period and the second predetermined period in this embodiment are each 30 seconds. However, the present invention is not limited thereto. The first predetermined period and the second predetermined period can each be any period during which the acoustic sensor 60 can collect breathing sounds. For example, the first predetermined period and the second predetermined period may each be shorter than the constant period CP in the acoustic signal storage processing. In other words, the constant period CP in the acoustic signal storage processing may be longer than the first predetermined period and the second predetermined period. Furthermore, the first predetermined period and the second predetermined period may be different from each other.

[0087] The respiratory sound detection device 20 of this embodiment can display two detection results: a first detection result for a first predetermined period, and a second detection result for a second predetermined period that follows the first predetermined period. Therefore, by issuing a start command before an event such as a meal and an end command after the event, the detection results before and after the event can be compared. By comparing the detection results, it can be determined that the abnormal respiratory sound in the second detection result was caused by the event.

[0088] More specifically, if the pre-meal assessment result for subject 70 is normal, but the post-meal assessment result for subject 70 is abnormal, it can be determined that abnormal breathing occurred due to the meal. For example, it can be determined that there is a possibility that a foreign object caused by the meal remains in the upper part of subject 70's respiratory organs. In other words, according to this embodiment, a respiratory sound detection device 20 that detects abnormal breathing caused by events such as meals can be provided.

[0089] In this embodiment, the event is eating. However, the present invention is not limited to this. For example, the event may be exercise, bathing, or taking medication such as an asthma inhaler. A first determination process may be performed before these events, and a second determination process may be performed after these events. By using the respiratory sound determination device 20 of this embodiment for such purposes, it is possible not only to detect abnormal breathing caused by an event, but also to detect when abnormal breathing has disappeared due to an event such as light exercise or medication.

[0090] Referring to Figure 10 in conjunction with Figure 6, as mentioned above, the acoustic signal storage process displays a confirmation message on the display device 26 if the CP is unable to receive an acoustic signal SS for a certain period of time, or if the received acoustic signal SS does not contain a breathing sound. Furthermore, the acoustic signal storage process is executed on a different thread from the control process (see Figure 9), the first determination process (see Figure 11), and the second determination process (see Figure 12). Therefore, the acoustic signal storage process can display a confirmation message on the display device 26 at various timings.

[0091] For example, the acoustic signal storage process can display a confirmation message on the display device 26 both before the first determination process (first timing) and before the second determination process (second timing). Furthermore, the acoustic signal storage process can change the content of the confirmation message depending on the timing. For example, if the respiratory sound determination device 20 is unable to acquire respiratory sounds for a certain period of time after performing the first determination process but before being instructed to terminate, it can display information on the display device 26 indicating that respiratory sounds could not be acquired. Such a display allows the operator of the respiratory sound determination device 20 to become aware of various problems, such as a malfunction of the acoustic sensor 60, in a timely manner.

[0092] The respiratory sound detection device 20 of this embodiment can be modified in various ways in addition to the various modifications already described. Modifications of the respiratory sound detection device 20 will be described below.

[0093] Referring to Figure 1, according to this embodiment, the first and second judgment results are displayed on the display device 26, respectively. However, the present invention is not limited thereto. For example, if the second judgment result is "abnormal", the respiratory sound determination device 20 may emit a sound through its speaker.

[0094] Referring to Figure 8, the first predetermined period in this embodiment is a continuous period immediately following the operator of the respiratory sound determination device 20 giving a start command. However, the present invention is not limited thereto. For example, the first predetermined period may be a discontinuous period starting a few seconds after the operator gives the start command. The sum of two or more such discontinuous periods may constitute the first predetermined period. That is, the first predetermined period may be the period after the start command. Also, referring to Figure 9 in conjunction with Figure 6, the respiratory sound determination program 30 may start the acoustic signal recording process (S920) immediately after starting operation (i.e., before S910). Referring to Figure 8, in this case, the first predetermined period may include the period before the operator gives the start command, or it may consist only of the period before the operator gives the start command.

[0095] In this embodiment, the second predetermined period is a continuous period immediately preceding the operator giving the termination command. However, the present invention is not limited thereto. For example, the second predetermined period may be a discontinuous period ending a few seconds before the operator gives the termination command. The sum of two or more such discontinuous periods may constitute the second predetermined period. That is, the second predetermined period may be the period before the termination command is given. Furthermore, the second predetermined period may include the period after the operator gives the termination command, or it may consist only of the period after the operator gives the termination command.

[0096] Referring to Figure 1, as described above, the respiratory sound determination device 20 of this embodiment, after performing the second determination process, transmits the first determination result and the second determination result to a different management device 40, depending on the combination of the first determination result and the second determination result. However, the present invention is not limited thereto. For example, the respiratory sound determination device 20 may transmit the first determination result and the second determination result to the management device 40 regardless of the content of the first and second determination results. Also, the respiratory sound determination device 20 does not have to transmit the first determination result and the second determination result to the management device 40.

[0097] The respiratory sound detection device 20 of this embodiment performs respiratory sound detection only in the first and second detection processes. That is, according to the respiratory sound detection device 20 of this embodiment, two detection results are obtained only for two predetermined periods, the first and second, which are separated from each other by an event. However, the present invention is not limited thereto. For example, the respiratory sound detection device 20 may perform respiratory sound detection in a third predetermined period in the middle of an event, in addition to the first and second detection processes.

[0098] Referring to Figures 11 and 12, the respiratory sound detection device 20 of this embodiment determines whether respiratory sounds are normal or abnormal using an abnormality discrimination model. The abnormality discrimination model of this embodiment is a neural network trained using deep learning. According to this embodiment, for example, it is possible to determine whether respiratory sounds are normal or abnormal with an accuracy of 80% or more. However, the present invention is not limited thereto, and the method for determining whether respiratory sounds are normal or abnormal is not particularly limited.

[0099] The anomaly detection model of this embodiment (see Figure 13) will be described below.

[0100] Referring to Figure 13 in conjunction with Figures 6, 11, and 12, the respiratory sound detection device 20 applies a short-time Fourier transform (STFT) to the acoustic signal SS for a first predetermined period or a second predetermined period to create a spectrogram (S1110 in Figure 11, S1210 in Figure 12, and S1310 in Figure 13). The respiratory sound detection device 20 uses the created spectrogram as an input layer to obtain the output layer of the abnormality discrimination model (S1130 in Figure 11, S1230 in Figure 12, and S1320 in Figure 13).

[0101] The spectrogram size of this embodiment is 249 × 129 pixels and has 1 channel. The anomaly detection model of this embodiment is a convolutional neural network (CNN). For example, the anomaly detection model of this embodiment is a machine-trained neural network having the layer structure shown in Figure 14. The output layer of the anomaly detection model is the probability that the respiratory sound is abnormal, and is a numerical value between 0 and 100%. However, the present invention is not limited to this. The anomaly detection model may be of any type. The input and output layers of the anomaly detection model can be modified as needed. For example, the input layer may be an inverted spectrogram.

[0102] The anomaly detection model of this embodiment can be generated, for example, as shown in Figure 15.

[0103] First, acoustic signals SS during meals are collected, creating numerous acoustic signal files (S1510). Next, specialists such as speech-language pathologists listen to the acoustic signal files and extract abnormal breathing (S1520). More specifically, the start and end times of abnormal breathing sounds are recorded for each acoustic signal file. Based on these start and end times, numerous acoustic signal files are created, each containing sections where abnormal breathing occurs. Next, each acoustic signal file is subjected to a short-time Fourier transform to create numerous spectrograms for machine learning (S1530). Then, machine learning is performed using the created spectrograms as training data (S1540). As a result of this machine learning, an anomaly detection model is generated. Next, the generated anomaly detection model is converted into a lightweight anomaly detection model that can be installed on a mobile device using a converter.

[0104] The anomaly detection model of this embodiment can be generated as described above. However, the method for generating the anomaly detection model according to the present invention is not particularly limited.

[0105] For example, instead of a lightweight anomaly detection model that does not learn, an anomaly detection model that continues to learn may be installed on the mobile device. In this case, referring to Figure 6, the respiratory sound detection program 30 may retain the acoustic data 32 without deleting it when it terminates its operation. That is, it may continue to accumulate the acoustic data 32. By determining whether the respiratory sounds in the large amount of accumulated acoustic data 32 are normal or abnormal, the accuracy of the anomaly detection model can be improved.

[0106] The respiratory sound determination system 10 of this embodiment (see Figure 1) can be modified as described below.

[0107] Referring to Figure 16, the modified respiratory sound detection system 10A comprises a respiratory sound detection device 20 and two or more acoustic sensors 60. The acoustic sensors 60 are attached to two or more subjects 70. When each acoustic sensor 60 starts operating, it collects ambient sounds, converts them into acoustic signals SS, and transmits them externally. The respiratory sound detection device 20 is paired with each of the acoustic sensors 60. The respiratory sound detection device 20 receives the acoustic signals SS transmitted by the two or more acoustic sensors 60.

[0108] Referring to Figure 16 in conjunction with Figure 17, the respiratory sound detection device 20 of this modified example performs control processing (see Figure 9), acoustic signal storage processing (see Figure 10), first determination processing (see Figure 11), and second determination processing (see Figure 12) for each of the paired acoustic sensors 60. For example, the respiratory sound detection device 20 can process two or more acoustic sensors 60 selected from a sensor list screen (not shown) displayed on the display device 26 in parallel. That is, the respiratory sound detection device 20 of this modified example can perform first determination processing for respiratory sounds in two or more acoustic signals SS whose first predetermined periods overlap, and can perform second determination processing for respiratory sounds in two or more acoustic signals SS whose second predetermined periods overlap.

[0109] When pairing the respiratory sound detection device 20 of this modified example with each of the acoustic sensors 60, the respiratory sound detection device 20 may read the two-dimensional code of the tag 69 (see Figure 2) attached to the base member 67 (see Figure 2) and store the identification name of the acoustic sensor 60 in association with the acoustic sensor 60. By storing the identification name of the acoustic sensor 60 in association with the acoustic sensor 60, for example, when the respiratory sound detection device 20 displays the first determination result on the display device 26, the identification name of the acoustic sensor 60 can also be displayed.

[0110] According to this modified example, abnormal breathing in two or more subjects 70 can be detected by a single respiratory sound detection device 20. On the other hand, the respiratory sound detection system 10A may also include two or more respiratory sound detection devices 20 and two or more acoustic sensors 60. In this case, the acoustic sensors 60 paired with the two or more respiratory sound detection devices 20 may be different from each other or may overlap. If the acoustic sensors 60 paired with the respiratory sound detection devices 20 overlap, abnormal breathing in one subject 70 can be detected by two or more respiratory sound detection devices 20. Therefore, the possibility of missing abnormal breathing can be reduced.

[0111] As can be understood from the above explanation, this modified version provides a monitoring method for two or more subjects 70 using the respiratory sound detection system 10A. Each subject 70 is fitted with an acoustic sensor 60. Before all subjects 70 perform a predetermined event such as eating, the acoustic sensor 60 is activated and a start command is sent to the respiratory sound detection device 20. After all subjects 70 have performed the predetermined event, a stop command is sent to the respiratory sound detection device 20. With this monitoring method, 70 target participants It allows for efficient caregiving. [Explanation of Symbols]

[0112] 10,10A Respiratory Sound Detection System 20 Breathing sound determination device 21 Processing Unit 22 Main storage 23 Auxiliary storage device 25 Input devices 26 Display device 28 Antennas 29 Additional antennas 30. Respiratory Sound Analysis Program 32 Audio Data 32S Start Record 32E End Record 322 Start time 324 End time 328 partial audio data 40 Management device 60 Acoustic Sensors 62 Sensor body 622 External Microphone 624 Locked part 63 switches 64 lamps 65 connectors 66 Contact area 662 recess 664 Sound Collection Sheet 67 Base member 672 Bottom 674 Lock section 68 Passing hole 69 tags 70 Target Persons SS acoustic signal

Claims

1. A respiratory sound detection device that determines respiratory sounds in an acoustic signal received from an external source, The aforementioned respiratory sound determination device is equipped with a display device, The respiratory sound detection device performs a first detection process when it is instructed to start, and performs a second detection process when it is instructed to end after it has been instructed to start. In the first determination process, the respiratory sound determination device displays on the display device a first determination result indicating whether the respiratory sound during a first predetermined period is normal or abnormal. In the second determination process, the respiratory sound determination device displays on the display device a second determination result indicating whether the respiratory sound during the second predetermined period is normal or abnormal. The first predetermined period and the second predetermined period are two periods separated from each other by an event, and the first predetermined period is a period prior to the second predetermined period. Breathing sound determination device.

2. A respiratory sound determination device according to claim 1, After performing the second determination process, the respiratory sound determination device transmits the first determination result and the second determination result to a different management device, depending on the combination of the first determination result and the second determination result. Breathing sound determination device.

3. A respiratory sound determination device according to claim 1, The aforementioned respiratory sound detection device is a mobile terminal, The display device is the screen of the mobile terminal. Breathing sound determination device.

4. A respiratory sound determination device according to any one of claims 1 to 3, If the respiratory sound detection device is unable to acquire the respiratory sound for a certain period of time after performing the first detection process but before being instructed to terminate, it will display information on the display device indicating that the respiratory sound could not be acquired. Breathing sound determination device.

5. A respiratory sound determination device according to any one of claims 1 to 3, The respiratory sound determination device can perform the first determination process for the respiratory sounds in two or more acoustic signals whose first predetermined periods overlap, and can perform the second determination process for the respiratory sounds in two or more acoustic signals whose second predetermined periods overlap. Breathing sound determination device.

6. This is a program for causing a mobile terminal to function as a respiratory sound determination device according to any one of claims 1 to 3. A program for analyzing respiratory sounds.

7. A respiratory sound determination system comprising a respiratory sound determination device according to any one of claims 1 to 3 and two or more acoustic sensors, Each of the aforementioned acoustic sensors, upon starting operation, collects ambient sounds, converts them into acoustic signals, and transmits them externally. The respiratory sound detection device receives the acoustic signals transmitted by two or more acoustic sensors. Respiratory sound detection system.

Images