JP7260411B2 - Acoustic monitoring device - Google Patents

Description

The present invention relates to acoustic monitoring devices.

2. Description of the Related Art Conventionally, there is an acoustic monitoring device that stores acoustic signals in a storage area and uses them for specific information processing. It is known that in this acoustic monitoring device, it is necessary to keep user data stored in the storage area confidential from the viewpoint of privacy protection and security. For example, Patent Literature 1 discloses information processing that protects user data by providing a storage unit that stores voice data of a user and an erasing unit that deletes the voice data stored in the storage unit after a predetermined time has elapsed. A device is described.

JP 2017-182075 A

In the device described in Patent Document 1, the monitored object is the voice of a specific user who uses the device, not the acoustic signal in a public space where an unspecified number of people gather. However, when implementing acoustic monitoring, even if the main object of monitoring is the operating sound of equipment installed in public spaces, acoustic signals related to privacy, such as the voices of the general public, are inevitably included. there is a possibility. For this reason, there has been a problem in achieving compatibility between protection of the privacy of the general public and sound monitoring in public spaces.

The present invention focuses on the above problem, and an object of the present invention is to provide an acoustic monitoring apparatus capable of performing acoustic monitoring in a public space while protecting the privacy of the general public.

In order to solve the above problems, one typical acoustic monitoring device of the present invention is characterized by decomposing an acoustic signal temporarily stored in a storage area into a first acoustic feature quantity and a second acoustic feature quantity. a quantity decomposition unit, a feature selection unit that selects one of the first acoustic feature and the second acoustic feature, and the first acoustic feature selected by the feature selection unit a feature quantity discarding unit that discards the quantity or the second acoustic feature quantity and stores a residual acoustic feature quantity composed of the first acoustic feature quantity or the second acoustic feature quantity in the storage area. ing.

ADVANTAGE OF THE INVENTION According to this invention, the acoustic monitoring apparatus which can implement acoustic monitoring in a public space can be provided, protecting the privacy of unspecified many people.
Problems, configurations, and effects other than those described above will be clarified by the description in the following detailed description of the invention.

The figure which shows the whole structure of the acoustic monitoring apparatus which concerns on embodiment of this invention. FIG. 2 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of the acoustic monitoring device according to the first embodiment of the present invention; The figure which shows the structure of the feature-value decomposition|disassembly part of the acoustic monitoring apparatus which concerns on Example 2 of this invention, a feature-value selection part, and a feature-value discarding part. The figure which shows the structure of the feature-value decomposition|disassembly part of the acoustic monitoring apparatus based on Example 3 of this invention, a feature-value selection part, and a feature-value discarding part. The figure which shows the structure of the feature-value decomposition|disassembly part, the feature-value selection part, and the feature-value discarding part of the acoustic monitoring apparatus which concerns on Example 4 of this invention. The figure which shows the example of the frequency band decomposition|disassembly in the instantaneous amplitude component band decomposition|disassembly part with which the feature-value decomposition|disassembly part of the acoustic monitoring apparatus which concerns on Example 4 of this invention was equipped. The figure which shows the structure of the feature-value decomposition|disassembly part of the acoustic monitoring apparatus which concerns on Example 5 of this invention, a feature-value selection part, and a feature-value discarding part. FIG. 10 is a graph showing an example of a squared instantaneous amplitude component and its time average component in an average fluctuation component decomposing unit provided in a feature amount decomposing unit of the acoustic monitoring device according to Example 5 of the present invention; FIG. The figure which shows the structure of the feature-value decomposition|disassembly part of the acoustic monitoring apparatus which concerns on Example 6 of this invention, a feature-value selection part, and a feature-value discarding part. FIG. 10 is a graph showing an example of instantaneous phase components and their average gradient components in the gradient fluctuation component decomposition unit provided in the feature amount decomposition unit of the acoustic monitoring device according to the sixth embodiment of the present invention; FIG. 11 is a flow chart showing the processing of the feature quantity selection unit of the sound monitoring device according to Embodiment 7 of the present invention; FIG. FIG. 11 is a diagram showing an example of generation probabilities of a random binary number generator provided in a feature selector of the acoustic monitoring apparatus according to Embodiments 7 and 8 of the present invention; FIG. 11 is a flow chart showing processing of a feature selection unit of the sound monitoring device according to the eighth embodiment of the present invention; FIG. FIG. 11 is a diagram showing entropy of residual acoustic features obtained by the processing of the feature selection unit of the acoustic monitoring device according to the eighth embodiment of the present invention; The figure which shows the outline|summary of the sound-monitoring apparatus with which the public space where unspecified many people gather. The figure which shows the structure of a microcomputer.

Hereinafter, embodiments of the present invention will be described with reference to the drawings. It should be noted that the present invention is not limited by this embodiment. Moreover, in the description of the drawings, the same parts are denoted by the same reference numerals.

FIG. 15 shows a sound monitoring device installed in a public space where many and unspecified people gather. The acoustic monitoring device 1 converts an acoustic signal converted into an electric signal 3 using an electroacoustic transducer 2 into a digital signal 5 using an analog-to-digital converter 4, and stores the digital signal 5 in a storage area 6. . When using such an acoustic monitoring device to monitor an acoustic signal in a public space where an unspecified number of people 7 gather, for example, in a train cabin space 8, the electroacoustic transducer 2 receives noise in addition to the operating sound 9 of the train. A sound signal 10 related to privacy, such as a voice uttered by a specific large number of people 7, arrives. Therefore, there is a problem of compatibility between protection of the privacy of the general public 7 and acoustic monitoring in public spaces.

FIG. 1 shows the overall configuration of an acoustic monitoring device according to an embodiment of the present invention. The acoustic monitoring apparatus of this embodiment has a feature decomposing unit 15 that decomposes the acoustic signal 11 temporarily stored in the storage area 6 into a first acoustic feature 12 and a second acoustic feature 13 . Then, one of the decomposed first acoustic feature quantity 12 and the second acoustic feature quantity 13 is selected by the feature quantity selection unit 17 . Then, the first acoustic feature quantity 12 or the second acoustic feature quantity 13 selected 16 by the feature quantity selection unit 17 is discarded, and the acoustic feature quantity not discarded is stored in the storage area 6 as the residual acoustic feature quantity 14. be done.

As for the residual acoustic feature amount, since part of the information is deleted from the temporarily stored acoustic signal, the acoustic signal 10 related to privacy such as the voice of the general public 7 is inevitably mixed into the acoustic signal 11. Even so, it is difficult to restore the acoustic signal 10 related to privacy such as voice from the residual acoustic feature amount 14 . On the other hand, since it is possible to carry out information processing such as diagnosing the operation status of the device based on the residual acoustic feature amount, the acoustic monitoring device according to the present embodiment protects the privacy of the general public 7 and can be used in public spaces. Acoustic monitoring can be performed.

A public space acoustic monitoring functional component 19 comprising the feature quantity decomposition section 15, the feature quantity selection section 17 and the feature quantity discarding section 18 can be realized by, for example, a microcomputer shown in FIG. In the example of FIG. 16, the microcomputer comprises a CPU 51, a ROM 52, a RAM 53 and an input/output port 54, which can be connected to each other via a bus line 55.

For example, the ROM 52 can store a program that causes the CPU 51 to operate as the feature quantity decomposition unit 15 , the feature quantity selection unit 17 and the feature quantity discarding unit 18 . The RAM 53 can temporarily store the acoustic signal 11 , the first acoustic feature amount 12 and the second acoustic feature amount 13 input from the storage area 6 . Further, the input/output port 54 may input the acoustic signal 11 from the storage area 6 and output the residual acoustic feature quantity 14 to the storage area 6 .
Hereinafter, each embodiment of the present invention will be described in order with reference to FIGS. 2 to 14. FIG.

FIG. 2 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of the acoustic monitoring device according to the first embodiment of the present invention. The acoustic monitoring apparatus of the present embodiment, as a feature quantity decomposing unit, decomposes an acoustic signal 11 into a first acoustic feature quantity 20 consisting of amplitude frequency components and a second acoustic feature quantity 21 consisting of phase frequency components. An amplitude phase frequency decomposition unit 22 is provided for performing frequency analysis.

ここでΩは角周波数である。 That is, assuming that the digitized acoustic signal 11 is x(n), the amplitude phase frequency decomposition unit 22 converts the acoustic signal 11 into a first acoustic feature value |X(e ^−iΩ )| and the second acoustic feature quantity θ(e ^−iΩ ) consisting of frequency components as follows.

where Ω is the angular frequency.

For this reason, in the acoustic monitoring device according to the present embodiment, even if an acoustic signal related to privacy such as voices of an unspecified number of people is inevitably mixed in the acoustic signal 11, the original acoustic signal 11 is set to the amplitude frequency It is difficult to recover from the residual acoustic features 14 that consist of the first acoustic features 20 consisting of components or the second acoustic features 21 consisting of phase frequency components.

On the other hand, if the amplitude frequency component or the phase frequency component is used, information processing such as equipment condition diagnosis is possible. For example, if the state of the device reflects the amplitude component or phase component of a specific frequency, either the amplitude component or phase component of the specific frequency reflected by the state of the device is stored in the storage area, and the other is discarded. Just do it. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 3 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of the acoustic monitoring device according to the second embodiment of the present invention. The acoustic monitoring apparatus of this embodiment, as a feature quantity decomposing unit, decomposes the acoustic signal 11 into a first acoustic feature quantity 23 consisting of an instantaneous amplitude component and a second acoustic feature quantity 24 consisting of an instantaneous phase component. An amplitude phase decomposition section 25 is provided.

ここでフラクトゥールのＲは解析信号の実部を抽出する演算子である。 That is, assuming that the digitized acoustic signal 11 is x(n), the instantaneous amplitude phase decomposition unit 25 divides the acoustic signal 11 into a first acoustic feature |Z(n)| consisting of an instantaneous amplitude component and an instantaneous phase component and a second acoustic feature quantity θ(n) consisting of

Here, R in Fractur is an operator that extracts the real part of the analytic signal.

For this reason, in the acoustic monitoring device according to the present embodiment, even if an acoustic signal related to privacy such as voices of an unspecified number of people is inevitably mixed in the acoustic signal 11, the original acoustic signal 11 is not affected by the instantaneous amplitude. It is difficult to restore from the residual acoustic feature quantity 14 consisting of the first acoustic feature quantity 23 consisting of components or the second acoustic feature quantity 24 consisting of instantaneous phase components.

On the other hand, if the instantaneous amplitude component or the instantaneous phase component is used, information processing such as equipment condition diagnosis is possible. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 4 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of an acoustic monitoring device according to Example 3 of the present invention. The acoustic monitoring device of the present embodiment includes, as a feature amount decomposition unit, a narrowband decomposition unit 27 that decomposes the acoustic signal 11 into a plurality of narrowband acoustic signals 26 having a specific frequency band, and the narrowband acoustic signals 26 are A plurality of instantaneous amplitude/phase decomposition units 25 are provided for decomposing into a first acoustic feature quantity 23 consisting of instantaneous amplitude components and a second acoustic feature quantity 24 consisting of instantaneous phase components.

That is, assuming that the digitized acoustic signal 11 is x(n), the feature quantity decomposing unit of the present embodiment converts the acoustic signal 11 into a first acoustic feature quantity | It is decomposed into Z _m (n)| and a second acoustic feature quantity θ _m (n) consisting of an instantaneous phase component having a specific frequency band m as follows.

For this reason, in the acoustic monitoring device according to the present embodiment, even if an acoustic signal related to privacy such as voices of an unspecified number of people is inevitably mixed in the acoustic signal 11, the original acoustic signal 11 is not affected by the instantaneous amplitude. It is difficult to restore from the residual acoustic feature quantity 14 consisting of the first acoustic feature quantity 23 consisting of components or the second acoustic feature quantity 24 consisting of instantaneous phase components.

On the other hand, if an instantaneous amplitude component or instantaneous phase component having a specific frequency band m is used, information processing such as equipment state diagnosis is possible. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 5 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of an acoustic monitoring device according to Example 4 of the present invention. The acoustic monitoring apparatus of the present embodiment includes, as a feature decomposing unit, a narrowband decomposing unit 27 that decomposes the acoustic signal 11 into a plurality of narrowband acoustic signals 26 having a specific frequency band, and an instantaneous a plurality of instantaneous amplitude component extraction units 28 for extracting the amplitude component 29; A plurality of instantaneous amplitude component band decomposing units 32 are provided for decomposing into a second acoustic feature quantity 31 consisting of a plurality of components.

That is, as shown in FIG. 6, the feature quantity decomposing unit of the present embodiment divides the instantaneous amplitude component into a first acoustic feature quantity consisting of a first band-specific instantaneous amplitude component 33 and a second band-specific instantaneous amplitude component 34 second acoustic features. Then, only the residual acoustic feature amount composed of any of them is stored in the storage area.

Therefore, in the acoustic monitoring apparatus according to the present embodiment, even if the acoustic signal 11 inevitably contains an acoustic signal related to privacy such as voices of an unspecified number of people, the original acoustic signal 11 is It is difficult to restore from the residual acoustic feature quantity 14 composed of the first acoustic feature quantity 30 consisting of the instantaneous amplitude component by band or the second acoustic feature quantity 31 consisting of the instantaneous amplitude component by band of .

Furthermore, the first band-specific instantaneous amplitude component 33 and the second band-specific instantaneous amplitude component 34 are appropriately set for phonological information uttered by an unspecified number of people transmitted as instantaneous amplitude components in a specific band. By doing so, the corresponding phoneme information or the like can be discarded.

On the other hand, if the first band-specific instantaneous amplitude component 33 or the second band-specific instantaneous amplitude component 34 included in the residual acoustic feature quantity 14 is used, information processing such as device status diagnosis is possible. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 7 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of the acoustic monitoring device according to the fifth embodiment of the present invention. The acoustic monitoring apparatus of the present embodiment includes, as a feature decomposing unit, a narrow band decomposing unit 27 that decomposes the acoustic signal 11 into a plurality of narrow band acoustic signals 26 having a specific frequency band; a plurality of instantaneous square amplitude component extraction units 36 for extracting the instantaneous square amplitude component 35; and a second acoustic feature quantity 38 consisting of a time-varying component obtained by subtracting the time-average component from the instantaneous amplitude component, and a plurality of mean-fluctuation-component decomposing units 39 for decomposing.

That is, as shown in FIG. 8, the feature quantity decomposing unit of this embodiment has a first acoustic feature quantity 37 consisting of a time-averaged component of the instantaneous squared amplitude component 35, and the first acoustic feature quantity 37 from the instantaneous squared amplitude component 35. It decomposes into a second acoustic feature quantity 38 consisting of a time-varying component from which the feature quantity 37 is subtracted. Then, only the residual acoustic feature amount composed of any of them is stored in the storage area.

For this reason, in the acoustic monitoring device according to the present embodiment, even if an acoustic signal related to privacy such as voices of an unspecified number of people is unavoidably mixed in the acoustic signal 11, the original acoustic signal 11 is time-averaged. It is difficult to recover from the residual acoustic features 14, which consist of the first acoustic features 37 consisting of components or the second acoustic features 38 consisting of time-varying components.

On the other hand, if the first acoustic feature quantity 37 consisting of the time-averaged component or the second acoustic feature quantity 38 consisting of the time-varying component included in the residual acoustic feature quantity 14 is used, information processing such as equipment condition diagnosis is possible. be. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

Note that the interval of time averaging can be arbitrarily set in the acoustic monitoring device.

FIG. 9 is a diagram showing configurations of a feature decomposing unit, a feature selecting unit, and a feature discarding unit of an acoustic monitoring apparatus according to Example 6 of the present invention. The acoustic monitoring apparatus of the present embodiment includes, as a feature decomposing unit, a narrowband decomposing unit 27 that decomposes the acoustic signal 11 into a plurality of narrowband acoustic signals 26 having a specific frequency band, and an instantaneous a plurality of instantaneous phase component extracting units 41 for extracting the phase component 40; a first acoustic feature quantity 42 consisting of an average gradient component of the instantaneous phase component 40; and a second acoustic feature amount 43 consisting of time-varying components from which the tilt component has been subtracted, and a plurality of tilt fluctuation component decomposing units 44 for decomposing.

That is, as shown in FIG. 10, the feature quantity decomposing unit of the present embodiment has a first acoustic feature quantity 42 consisting of the average gradient component of the instantaneous phase component 40, and the first acoustic feature quantity from the instantaneous phase component 40. 42 is decomposed into a second acoustic feature amount 43 consisting of time-varying components. Then, only the residual acoustic feature amount composed of any of them is stored in the storage area.

Therefore, in the acoustic monitoring device according to the present embodiment, even if the acoustic signal 11 inevitably includes an acoustic signal related to privacy such as voices of an unspecified number of people, the original acoustic signal 11 is averaged. It is difficult to restore from the first acoustic feature quantity 42 consisting of components or the second acoustic feature quantity 43 consisting of time-varying components.

On the other hand, if the first acoustic feature quantity 42 consisting of the average gradient component or the second acoustic feature quantity 43 consisting of the time-varying component included in the residual acoustic feature quantity 14 is used, information processing such as equipment condition diagnosis is possible. be. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 11 is a flow chart showing the processing of the feature selector of the acoustic monitoring device according to the seventh embodiment of the present invention. According to this, the selection 16 of the acoustic feature quantity selector for selecting either the first acoustic feature quantity 12 or the second acoustic feature quantity 13 shown in FIG. determined by the instrument output.

That is, as shown in FIG. 12, the 0 output 48 or 1 output 49 of the random binary number generator is determined by the specified probabilities 1-p and p, respectively. Here, p is an arbitrary real number from 0 to 1, and can be arbitrarily set in the acoustic monitoring device.

Therefore, in the acoustic monitoring apparatus according to the present embodiment, the residual acoustic feature quantity 14 stored in the storage area 6 is probabilistic. is unavoidably mixed, it is difficult to restore the original acoustic signal 11 from the first acoustic feature quantity 12 or the second acoustic feature quantity 13 .

On the other hand, if it is known that the residual acoustic feature quantity 14 stored in the storage area 6 is based on the specified probability, the first acoustic feature quantity 12 or the second acoustic feature quantity included in the residual acoustic feature quantity 14 Information processing such as equipment condition diagnosis using the quantity 13 is possible. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

FIG. 13 is a flow chart showing the processing of the feature selector of the acoustic monitoring device according to the eighth embodiment of the present invention. According to this, the selection 16 of the acoustic feature quantity selector that selects either the first acoustic feature quantity 12 or the second acoustic feature quantity 13 is the output of the random binary number generator based on the 50% generation probability determined by

That is, as shown in FIG. 14, the entropy H is maximized when p=0.5 with a 50% generation probability.

Therefore, in the acoustic monitoring apparatus according to the present embodiment, the uncertainty of the residual acoustic feature quantity 14 stored in the storage area 6 is maximized, and if the acoustic signal 11 is assumed to be the voice of an unspecified number of people, etc. Even if signals are unavoidably mixed, it becomes difficult to restore the original acoustic signal 11 from the first acoustic feature amount 12 or the second acoustic feature amount 13 .

On the other hand, if the residual acoustic feature quantity 14 stored in the storage area 6 is analyzed based on the conditional probability, the first acoustic feature quantity 12 or the second acoustic feature quantity 13 included in the residual acoustic feature quantity 14 can be determined as Information processing such as condition diagnosis of the equipment used is possible. Therefore, it is possible to provide an acoustic monitoring device capable of performing acoustic monitoring in a public space while protecting the privacy of an unspecified number of people.

Although the embodiments of the present invention have been described above, the present invention is not limited to the above-described embodiments, and various modifications are possible without departing from the gist of the present invention.

1 Acoustic monitoring device 2 Electro-acoustic transducer 3 Electric signal 4 Analog-to-digital converter 5 Digital signal 6 Storage area 7 Unspecified number of people 8 Train cabin space 9 Operating sound of train 10 Acoustic signal related to privacy such as voice 11 Acoustic signal 12 First acoustic feature quantity 13 Second acoustic feature quantity 14 Residual acoustic feature quantity 15 Feature quantity decomposition unit 16 Selection 17 Feature quantity selection unit 18 Feature quantity discarding unit 19 Public space sound monitoring functional component 20 First acoustic feature consisting of amplitude frequency components Acoustic feature quantity 21 Second acoustic feature quantity 22 consisting of phase frequency components Amplitude phase frequency decomposition unit 23 First acoustic feature quantity 24 consisting of instantaneous amplitude components Second acoustic feature quantity 25 consisting of instantaneous phase components Instantaneous amplitude phase Decomposing unit 26 Narrowband acoustic signal 27 Narrowband decomposing unit 28 Instantaneous amplitude component extracting unit 29 Instantaneous amplitude component 30 First acoustic feature amount 31 consisting of first instantaneous amplitude component by band 31 Consisting of second instantaneous amplitude component by band Second acoustic feature quantity 32 Instantaneous amplitude component band resolution unit 33 First instantaneous amplitude component by band 34 Second instantaneous amplitude component by band 35 Instantaneous square amplitude component 36 Instantaneous square amplitude component extraction unit 37 Time of square instantaneous amplitude component A first acoustic feature quantity 38 consisting of an average component A second acoustic feature quantity 39 consisting of a time variation component obtained by subtracting the time average component from the squared instantaneous amplitude component Average variation component decomposing unit 40 Instantaneous phase component 41 Instantaneous phase component extracting unit 42 A first acoustic feature quantity 43 consisting of an average gradient component of the instantaneous phase component A second acoustic feature quantity 44 consisting of a time-varying component obtained by subtracting the mean gradient component from the instantaneous phase component Tilt fluctuation component decomposing unit 48 Random binary number generator 0 output 49 Random binary generator 1 output 51 CPU
52 ROMs
53 RAM
54 input/output port 55 bus line

Claims (9)

An acoustic monitoring device comprising an electroacoustic transducer that converts an acoustic signal into an electrical signal and a storage area that stores the acoustic signal,
a feature decomposing unit that decomposes the acoustic signal temporarily stored in the storage area into a first acoustic feature and a second acoustic feature;
a feature selector that selects one of the first acoustic feature and the second acoustic feature;
discarding the first acoustic feature quantity or the second acoustic feature quantity selected by the feature quantity selection unit, and a residual acoustic feature composed of the first acoustic feature quantity or the second acoustic feature quantity; and a feature quantity discarding unit that stores the quantity in the storage area. 2. The acoustic monitoring device according to claim 1, wherein the feature quantity decomposition unit divides the acoustic signal into the first acoustic feature quantity consisting of amplitude frequency components and the second acoustic feature quantity consisting of phase frequency components. An acoustic monitoring device comprising an amplitude-phase-frequency resolving unit that decomposes into 2. The acoustic monitoring apparatus according to claim 1, wherein the feature quantity decomposition unit divides the acoustic signal into the first acoustic feature quantity consisting of an instantaneous amplitude component and the second acoustic feature quantity consisting of an instantaneous phase component. An acoustic monitoring device characterized by comprising an instantaneous amplitude-phase resolving unit that decomposes into . 2. The acoustic monitoring device according to claim 1, wherein the feature amount decomposition unit
a narrowband decomposition unit that decomposes the acoustic signal into a plurality of narrowband acoustic signals having specific frequency bands;
a plurality of instantaneous amplitude-phase decomposing units for decomposing the narrowband acoustic signal into the first acoustic feature amount composed of instantaneous amplitude components and the second acoustic feature amount composed of instantaneous phase components. An acoustic monitoring device characterized by: 2. The acoustic monitoring device according to claim 1, wherein the feature amount decomposition unit
a narrowband decomposition unit that decomposes the acoustic signal into a plurality of narrowband acoustic signals having specific frequency bands;
a plurality of instantaneous amplitude component extraction units for extracting instantaneous amplitude components from the narrowband acoustic signal;
A plurality of instantaneous amplitude components for decomposing the instantaneous amplitude component into the first acoustic feature quantity consisting of first band-specific instantaneous amplitude components and the second acoustic feature quantity consisting of second band-specific instantaneous amplitude components. An acoustic monitoring device comprising: a band decomposing unit; 2. The acoustic monitoring device according to claim 1, wherein the feature amount decomposition unit
a narrowband decomposition unit that decomposes the acoustic signal into a plurality of narrowband acoustic signals having specific frequency bands;
a plurality of square instantaneous amplitude component extraction units for extracting a square instantaneous amplitude component from the narrowband acoustic signal;
The squared instantaneous amplitude component is the first acoustic feature amount consisting of the time-averaged component of the squared instantaneous amplitude component, and the second sound consisting of the time-varying component obtained by subtracting the time-averaged component from the squared instantaneous amplitude component. and a plurality of average fluctuation component decomposition units that decompose into feature quantities. 2. The acoustic monitoring device according to claim 1, wherein the feature amount decomposition unit
a narrowband decomposition unit that decomposes the acoustic signal into a plurality of narrowband acoustic signals having specific frequency bands;
a plurality of instantaneous phase component extraction units for extracting instantaneous phase components from the narrowband acoustic signal;
The instantaneous phase components are the first acoustic feature quantity consisting of an average gradient component of the instantaneous phase component, and the second acoustic feature quantity consisting of a time-varying component obtained by subtracting the average gradient component from the instantaneous phase component. and a plurality of tilt fluctuation component decomposition units that decompose into . 8. The acoustic monitoring device according to any one of claims 1 to 7, wherein the feature selector includes a random binary number generator that generates a binary number with an arbitrary probability, and based on the binary number, An acoustic monitoring apparatus, wherein either one of the first acoustic feature quantity and the second acoustic feature quantity is selected. 9. The acoustic monitoring device of claim 8, wherein the random binary number generator has a binary number generation probability of 50%.

Images