Apnea episode determination device and apnea episode determination method

• US 9,629,582 B2
• Filed: 12/03/2012
• Issued: 04/25/2017
• Est. Priority Date: 03/27/2012
• Status: Active Grant

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Ishikawa, Chisato, Togawa, Taro, Otani, Takeshi, Suzuki, Masanao, Nishida, Asako
• Primary Examiner(s): JANG, CHRISTIAN YONGKYUN
• Application Number:
  US13/692,463
• Publication Number:
  US 20130261485A1
• Time in Patent Office: 1,604 Days
• Field of Search: 600529
• CPC Class Codes: A61B 5/4818 Sleep apnoea
• CPC Class Codes: A61B 5/7257 using Fourier transforms
• CPC Class Codes: A61B 7/003 Detecting lung or respiration noise

Abstract

An apnea episode determination device includes, a processor; and a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute, detecting a breathing segment and a midway segment from a sound signal during sleep, the breathing segment being considered to include a breathing sound, the midway segment existing in between the breathing segments; calculating an acoustic feature based on a background noise component and a signal component excluding the background noise component, which are included in the midway segment; and determining that the midway segment is an apnea episode when the acoustic feature meets a preset condition.

Claims

1. An apnea episode determination device comprising:

a processor; and

a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute,

dividing a sound signal from a subject during sleep into a plurality of frames, respective frames among the plurality of frames having a preset time length;

determining whether the respective frames include a breathing sound of the subject;

detecting, based on the determining, a breathing segment and a midway segment from the respective frames, the breathing segment having respective frames including the breathing sound, the midway segment having respective frames excluding the breathing sound and the midway segment existing between the breathing segments;

calculating a time length of the midway segment by counting a number of successive respective frames excluding the breathing sound in the midway segment;

calculating an acoustic feature based on a background noise component and a signal component excluding the background noise component, which are included in the midway segment; and

determining the midway segment is an apnea episode based on the time length of the midway segment and the acoustic feature, even when the background noise component fluctuates, when the acoustic feature meets a first preset condition and the time length of the midway segment is within a preset range,

wherein the acoustic feature is a signal-to-noise ratio in the midway segment calculated from one of a calculated sum, an average value and a weighted average of respective signal-to-noise ratios in the respective frames excluding the breathing sound included in the midway segment, and the first preset condition is when the signal-to-noise ratio is less than a preset threshold,

thereby the apnea episode is detected more accurately regardless of the background noise.

2. The device according to claim 1, further comprising:

the memory which stores a plurality of instructions, which when executed by the processor, cause the processor to further execute,

estimating a power of noise included in the sound signal in the respective frames among the plurality of frames having the preset time length,

wherein the signal-to-noise ratio in the midway segment is calculated based on the power of noise.

3. The device according to claim 2,

wherein the power of noise in one frame is estimated based on the power of noise in a last frame before the one frame and a sound volume of the sound signal included in the one frame when an amount of change in the volume of the sound signal included in the one frame is less than a preset value.

4. The device according to claim 2,

wherein the power of noise in one frame is estimated based on the power of noise in a last frame before the one frame and a sound volume of the sound signal included in the one frame when a ratio is less than a preset ratio, the ratio being a relationship between the volume of the sound signal included in the one frame and the power of noise of the last frame before the one frame.

5. The device according to claim 1, further comprising;

calculating an acoustic feature of at least one of the sound signals of the breathing segments immediately before and immediately after the midway segment,

wherein the midway segment is further determined as an apnea episode when the acoustic feature of the sound signal in the midway segment and the acoustic feature of at least one of the sound signals in the breathing segments immediately before and immediately after the midway segment meet preset conditions.

6. The device according to claim 5,

wherein a time length of at least one of the sound signals in the breathing segments is calculated immediately before and immediately after the midway segment, and

the midway segment is further determined as an apnea episode when the acoustic feature of the sound signal in the midway segment meets the first preset condition and the time length of the midway segment is longer than a preset threshold.

7. The device according to claim 5,

wherein the processor executes calculating an accumulated sound volume of at least one of the sound signals in the breathing segments immediately before and immediately after the midway segment, and the midway segment is further determined as an apnea episode when the acoustic feature of the sound signal in the midway segment meets the first preset condition and the accumulated sound volume is larger than a preset threshold.

8. The device according to claim 5,

wherein the processor executes calculating a difference in accumulated sound volume between the sound signals in the breathing segments immediately before and immediately after the midway segment, and the midway segment is further determined as an apnea episode when the acoustic feature of the sound signal in the midway segment meets the first preset condition and the difference in accumulated sound volume is less than a preset threshold.

9. The device according to claim 5,

wherein the processor executes calculating a spectrum distance between the sound signals in the breathing segments immediately before and immediately after the midway segment, and the midway segment is further determined as an apnea episode when the acoustic feature of the sound signal in the midway segment meets the first preset condition and the spectrum distance is less than a preset threshold.

10. The device according to claim 1,

wherein the calculating the time length of the midway segment is calculated by multiplying the number of successive respective frames excluding the breathing sound in the midway segment and the preset time length of the respective frames excluding the breathing sound in the midway segment.

11. An apnea episode determination method comprising:

dividing a sound signal from a subject during sleep into a plurality of frames, respective frames among the plurality of frames having a preset time length;

determining whether the respective frames include a breathing sound of the subject;

detecting, based on the determining, a breathing segment and a midway segment from the respective frames, the breathing segment having respective frames including the breathing sound, the midway segment having respective frames excluding the breathing sound and the midway segment existing between the breathing segments;

calculating a time length of the midway segment by counting a number of successive respective frames excluding the breathing sound in the midway segment;

calculating, by a computer processor, an acoustic feature based on a background noise component and a signal component excluding the background noise component, which are included in the midway segment; and

determining the midway segment is an apnea episode based on the time length of the midway segment and the acoustic feature, even when the background noise component fluctuates, when the acoustic feature meets a first preset condition and the time length of the midway segment is within a preset range,

wherein the acoustic feature is a signal-to-noise ratio in the midway segment, calculated from one of a calculated sum, an average value and a weighted average of respective signal-to-noise ratios in the respective frames excluding the breathing sound included in the midway segment and the first preset condition is when the signal-to-noise ratio is less than a preset threshold,

thereby the apnea episode is detected more accurately regardless of the background noise.

12. An apnea episode determination device comprising:

a processor; and

a memory which stores a plurality of instructions, which when executed by the processor, cause the processor to execute,

dividing a sound signal from a subject during sleep into a plurality of frames, respective frames among the plurality of frames having a preset time length;

determining whether the respective frames include a breathing sound of the subject;

detecting, based on the determining, a breathing segment and a midway segment from the respective frames, the breathing segment having respective frames including the breathing sound, the midway segment having respective frames excluding the breathing sound and the midway segment existing between the breathing segments;

calculating a time length of the midway segment by counting a number of successive respective frames excluding the breathing sound in the midway segment;

calculating an acoustic feature based on a background noise component and a signal component excluding the background noise component, which are included in the midway segment; and

determining the midway segment is an apnea episode based on the time length of the midway segment, the acoustic feature and a correlation between the sound signal in the midway segment and at least one of the sound signals in the breathing segments immediately before and immediately after the midway segment, even when the background noise component fluctuates, when the acoustic feature meets a first preset condition and the time length of the midway segment is within a preset range,

wherein the first preset condition is when the correlation is less than a preset threshold,

wherein the acoustic feature is a signal-to-noise ratio in the midway segment calculated from one of a calculated sum, an average value and a weighted average of respective signal-to-noise ratios in the respective frames excluding the breathing sound included in the midway segment, and the first preset condition is when the signal-to-noise ratio is less than a preset threshold,

thereby the apnea episode is detected more accurately regardless of the background noise.

13. The apnea episode determination device according to claim 12, wherein the correlation is calculated based on a difference in power spectrum between the sound signal in the midway segment and at least one of the sound signals in the breathing segments immediately before and immediately after the midway segment.