Device and method for detecting occurrence of wheeze

• US 9,113,780 B2
• Filed: 03/29/2012
• Issued: 08/25/2015
• Est. Priority Date: 03/29/2012
• Status: Active Grant

• Current Assignee: I-Shou University
• Original Assignee: I-Shou University
• Inventors: Chang, Gwo-Ching
• Primary Examiner(s): Zhang, Leshui
• Application Number:
  US13/433,729
• Publication Number:
  US 20130259240A1
• Time in Patent Office: 1,244 Days
• Field of Search: 381/56, 381/28, 381/60, 381/66, 381/91, 381/92, 381/94.1, 381/94.2, 381/94.3, 381/120, 381/122, 381/67, 600/529, 600/528, 600/533, 600/534, 600/543, 600/301, 600/586, 600/27, 128/204.23, 700/94
• US Class Current: 1/1
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise
• CPC Class Codes: H04R 2430/03 Synergistic effects of band splitting and sub-band processing
• CPC Class Codes: H04R 3/00 Circuits for transducers , loudspeakers or microphones

Abstract

A device for detecting occurrence of wheeze in a lung sound includes an acoustic processing unit, a filtering unit, and an identifying unit. The acoustic processing unit is for obtaining a lung sound and for generating a raw sound signal accordingly. The filtering unit is for filtering the raw sound signal to obtain a filtered signal. The filtered signal has a plurality of sections with respect to the time domain. The identifying unit is for computing an energy value of each of the sections of the filtered signal, and for determining that a wheeze occurs when the energy values of a part of the sections are larger than a predetermined value and duration of the part of the sections is longer than a predetermined observation time.

Claims

1. A device for detecting occurrence of wheeze in a lung sound, said device comprising:

an acoustic processing unit for obtaining a lung sound and for generating a raw sound signal in time domain accordingly;

a filtering unit for filtering the raw sound signal to obtain a filtered signal within a specific frequency bandwidth, the filtered signal having a plurality of sections with respect to the time domain; and

an identifying unit for computing an energy value of each of the plurality of the sections of the filtered signal, and for determining that a wheeze occurs in the lung sound when energy values of a part of the plurality of the sections are larger than a predetermined value and duration of said part of the sections is longer than a predetermined observation time in time domain;

wherein said filtering unit includes a plurality of biorthogonal wavelet filter banks connected in series, and each of the plurality of said biorthogonal wavelet filter banks of said filtering unit has a polyphase structure;

wherein, in the polyphase structure, each of the plurality of said biorthogonal wavelet filter banks receives an input signal, divides the input signal into a first part and a second part, and includes

a first filter set having a first high pass filter for filtering the first part of the input signal to output a first high-frequency signal, and a first low pass filter for filtering the first part of the input signal to output a first low-frequency signal, and

a second filter set having

a second high pass filter for filtering the second part of the input signal to output a second high-frequency signal, the second high-frequency signal being combined with the first high-frequency signal to form a high-frequency synthesis signal, and

a second low pass filter for filtering the second part of the input signal to output a second low-frequency signal, the second low-frequency signal being combined with the first low-frequency signal to form a low-frequency synthesis signal;

wherein, the raw sound signal from said acoustic processing unit serves as the input signal of a first one of the plurality of said biorthogonal wavelet filter banks of said filtering unit, one of the high-frequency and the low-frequency synthesis signals outputted by a last one of the plurality of said biorthogonal wavelet filter banks serves as the filtered signal, and one of the high-frequency and the low-frequency synthesis signals outputted by one of the plurality of said biorthogonal wavelet filter banks serves as the input signal of a next one of the plurality of said biorthogonal wavelet filter banks immediately following said one of the plurality of said biorthogonal wavelet filter banks.

2. The device as claimed in claim 1, wherein each of the plurality of said biorthogonal wavelet filter banks is a Cohen-Daubechies-Feauveru (CDF) 9/7 tap biorthogonal wavelet filter bank.

3. The device as claimed in claim 1, wherein each of the high-frequency and the low-frequency synthesis signals is obtained by distributed arithmetic, in which each of the first and the second high-frequency signals and the first and second low-frequency signals satisfies

4. The device as claimed in claim 1, wherein the specific frequency bandwidth of the filtered signal obtained by said filtering unit ranges from 250 Hertz to 500 Hertz.

5. The device as claimed in claim 1, wherein said identifying unit is operable to further compute relative energy values respectively by dividing the energy values of the plurality of the sections of the filtered signal by a given average energy value, and to determine that a wheeze occurs in the lung sound when the relative energy values associated with the energy values of the part of the plurality of the sections are larger than a threshold value and duration of the part of the plurality of the sections is longer than the predetermined observation time in the time domain.

6. The device as claimed in claim 1, further comprising an alert unit, wherein:

said identifying unit is operable to further generate an alert signal to said alert unit upon determination of the occurrence of the wheeze in the lung sound; and

said alert unit is operable, upon receipt of the alert signal, to generate an alert output to warn that a wheeze occurs in the lung sound.

7. The device as claimed in claim 6, wherein the alert output includes at least one of an audio output and a visual output.

8. The device as claimed in claim 1, wherein the predetermined observation time is 0.25 second.

9. The device as claimed in claim 1, further comprising a body at which said acoustic processing unit, said filtering unit and said identifying unit are disposed.

10. A method for detecting an occurrence of a wheeze in a lung sound, said method to be implemented by a detecting device, said method comprising the following steps of:

A) configuring the detecting device to obtain a lung sound and to generate a raw sound signal in time domain accordingly;

B) configuring the detecting device to filter the raw sound signal to obtain a filtered signal within a specific frequency bandwidth, the filtered signals having a plurality of sections with respect to the time domain;

C) configuring the detecting device to compute an energy value of each of the plurality of the sections of the filtered signal; and

D) configuring the detecting device to determine that the wheeze occurs in the lung sound when energy values of a part of the plurality of the sections are larger than a predetermined value and duration of said part of the sections is longer than a predetermined observation time in the time domain;

wherein the detecting device includes a plurality of biorthogonal wavelet filter banks connected in series, and each of the plurality of the biorthogonal wavelet filter banks has a polyphase structure;

wherein, in the polyphase structure, each of the plurality of said biorthogonal wavelet filter banks receives an input signal, divides the input signal into a first part and a second part, and includes

a first filter set having a first high pass filter for filtering the first part of the input signal to output a first high-frequency signal, and a first low pass filter for filtering the first part of the input signal to output a first low-frequency signal, and

a second filter set having

a second high pass filter for filtering the second part of the input signal to output a second high-frequency signal, the second high-frequency signal being combined with the first high-frequency signal to form a high-frequency synthesis signal, and

a second low pass filter for filtering the second part of the input signal to output a second low-frequency signal, the second low-frequency signal being combined with the first low-frequency signal to form a low-frequency synthesis signal;

wherein, the raw sound signal from said acoustic processing unit serves as the input signal of a first one of the plurality of said biorthogonal wavelet filter banks of said filtering unit, one of the high-frequency and the low-frequency synthesis signals outputted by a last one of the plurality of said biorthogonal wavelet filter banks serves as the filtered signal, and one of the high-frequency and the low-frequency synthesis signals outputted by one of the plurality of said biorthogonal wavelet filter banks serves as the input signal of a next one of the plurality of said biorthogonal wavelet filter banks immediately following said one of the plurality of said biorthogonal wavelet filter banks.

11. The method as claimed in claim 10, wherein, in step B), the specific frequency bandwidth ranges from 250 Hertz to 500 Hertz.

12. The method as claimed in claim 10, wherein, in step C), the detecting device is configured to further compute relative energy values respectively by dividing the energy values of the plurality of the sections of the filtered signal by a given average energy value, and to determine that a wheeze occurs in the lung sound when the relative energy values associated with the energy values of the part of the plurality of the sections are larger than a threshold value and a duration of the part of the plurality of the sections is longer than the predetermined observation time in the time domain.

13. The method as claimed in claim 10, further comprising, after step D), the following step of:

configuring the detecting device to generate an alert output to warn that the wheeze occurs in the lung sound.

14. The method as claimed in claim 13, wherein the alert output includes at least one of an audio output and a visual output.

15. The method as claimed in claim 10, wherein the predetermined observation time is 0.25 second.