JP6608229B2 - Information processing method and apparatus - Google Patents

Description

  The present invention relates to a technique for processing stuttering included in respiratory sounds.

  Abnormal sounds generated by contact with air, such as when respiratory secretions or blood accumulate in the trachea, bronchi, alveoli, lung cavities, etc., are called Russell sounds or rales (hereinafter referred to as ra sounds). The rales are classified as shown in FIG. Rashes are broadly classified into continuous continual rashes and intermittent intermittent raps, and when they are in contact with air, such as when breathing due to accumulation of sputum in the trachea, bronchi, alveoli, lung cavities, etc. Intermittent rales that emit are roaring. Intermittent rales have a short duration (10 ms or more and 25 ms or less) and a high frequency (250 Hz or more and 500 or less) bubbling sound, and a shorter duration (5 ms or less) and a higher frequency (500 Hz or more and 1000 Hz). And the following).

  Techniques for detecting rales contained in respiratory sounds have been developed in the past.

  For example, the local variance value in the frequency direction of the power spectrum is divided into a portion where the local variance value is larger and smaller than the threshold, and the inverse Fourier transform of the large portion is the waveform of the continuous rar sound, and the inverse Fourier transform of the small portion is the waveform of the normal breathing sound. (For example, Patent Document 1 below). However, the duration of the rarity is not evaluated.

  In addition, the Hilbert transform of the breathing sound is used to obtain an analysis signal, the waveform envelope, instantaneous frequency and instantaneous bandwidth are obtained, the peak is extracted from the envelope, and the extracted peak is based on the instantaneous frequency or instantaneous bandwidth of the extracted peak. There are also those that detect a rarity (for example, Patent Document 2 below). However, the duration of the rarity is not evaluated.

  Further, there is a technique for approximating the envelope of the frequency spectrum of respiratory sounds with a plurality of Gaussian waveforms, and classifying the respiratory sounds using the peak position and variance value of each Gaussian waveform as characteristic parameters of the respiratory sounds (for example, Patent Document 3 below). ). However, the duration of rales is not evaluated.

  Further, there is a technique for obtaining a ra sound by providing a threshold for a column of average values every 10 ms of a waveform obtained by converting the waveform of a high-frequency component of respiratory sound into an absolute value (for example, Non-Patent Document 1 below). Although the duration is evaluated, since analysis is performed by dividing the frame every 10 ms, a short phenomenon of 10 ms or less cannot be automatically detected. In addition, the frequency characteristics of rales are not evaluated.

  From the above, the conventional technology cannot detect the stuttering with high accuracy.

  In addition, doctors and nurses use a stethoscope to judge stuttering, but the auscultation requires advanced skills and experience. It is difficult to do. Therefore, there is a demand for a technique that assists caregivers, family members, and the like, such as confirming whether the sputum suction has been performed successfully, but such a technique does not exist.

JP 2004-357758 A JP 2009-106574 A JP 2013-123495 A

2011 Nihon University Faculty of Science and Technology Academic Lecture Proceedings “A Study on Detection of Abnormal Breathing Sound Using DSP”, Tetsuya Umezawa, Katsutoshi Saeki, Yoshifumi Sekine

  Therefore, the objective of this invention is providing the information processing technique for performing the effective assistance with respect to the person who performs sputum suction | inhalation as one side surface.

  The information processing method according to the present invention includes (A) a step of counting a first number of occurrences of stuttering included in a breathing sound of a predetermined period recorded at a first timing, and (B) at a second timing. The step of counting the second number of occurrences of stuttering included in the recorded breathing sounds for a predetermined period, and (C) the first timing and the second number of occurrences from the first number of occurrences and the second number of occurrences. A calculation step of calculating an evaluation value representing a symptom change with respect to timing.

  According to one aspect, it is possible to effectively assist a person performing sputum suction.

FIG. 1 is a diagram showing classification of rales. FIG. 2 is a diagram showing an outline of the embodiment of the present invention. FIG. 3 is a diagram illustrating an example of the information processing apparatus according to the present embodiment. FIG. 4 is a functional block diagram of the information processing apparatus according to the present embodiment. FIG. 5 is a functional block diagram of the count processing unit. FIG. 6 is a diagram showing a main processing flow according to the present embodiment. FIG. 7 is a diagram illustrating a processing flow of the counting process. FIGS. 8A to 8E are diagrams illustrating an example of a waveform change in the count process. FIGS. 9A and 9B are diagrams illustrating an example of a waveform change in the count process. FIG. 10 is a diagram illustrating an example of spectrum analysis. FIG. 11 is a diagram illustrating a processing flow of the stuttering determination processing. FIG. 12 is a diagram illustrating a display example of the processing result.

  An outline of the embodiment of the present invention is shown in FIG.

  The upper part of FIG. 2 schematically shows how the cared person's heel state changes from left to right as time elapses. In this example, when the amount of wrinkles gradually increases from a normal state and a certain amount of wrinkles remains, a caregiver or the like will suck the wrinkles. If sputum suction is performed properly, it returns to normal after sputum suction. In the present embodiment, for example, the information processing device determines whether or not sputum suction has been appropriately performed.

  That is, the information processing apparatus records the breathing sound immediately before the sputum suction and detects the stuttering and then counts the stuttering to obtain the count 1. Further, the information processing apparatus records the breathing sound immediately after the soot suction and detects the stuttering and then counts the stuttering to obtain the count 2. Then, the information processing apparatus evaluates the counts 1 and 2 that are the count results before and after the suction, and outputs the evaluation result of the wrinkle suction effect.

  This makes it possible for a caregiver or the like to easily determine whether or not the sputum has been properly sucked.

  For example, the information processing apparatus is a mobile terminal such as a smartphone to which an auscultation microphone as shown in FIG. 3 is connected, and performs a display as shown in FIG. 3 on the display device. In the example of FIG. 3, a button for instructing when the timing is before or after suction, a measurement start button, a measurement end button, a display column for measurement results (before suction), and a measurement result (after suction) ) Display column and improvement level display column. For example, after pressing the pre-suction button, pressing the measurement start button starts auscultation, and pressing the measurement end button ends the auscultation. Then, the number / unit time of stuttering before suction is displayed in the display column for the measurement result (before suction). In addition, after aspirating the cared person, press the button after suction, press the measurement start button to perform auscultation, and press the measurement end button to end the auscultation. Then, the number / unit time of stuttering after suction is displayed in the measurement result (after suction) display column, and the improvement based on the number of times before and after suction is calculated and displayed in the improvement display column. The

  A specific configuration of such an information processing apparatus will be described with reference to FIGS.

  As shown in FIG. 4, the information processing apparatus 100 according to the present embodiment is connected to an auscultation microphone 150, and includes an input unit 101, an input data storage unit 102, a count processing unit 103, and an evaluation unit 104. And an output unit 105.

  The input unit 101 digitizes the audio signal input from the auscultation microphone 150 and stores the digitized signal in the input data storage unit 102 or accepts a user input instruction from an input device such as a touch panel, for example, to other processing units. Execute the process.

  In response to an instruction from the input unit 101, the count processing unit 103 executes a process of counting stuttering from digital voice data (also referred to as a digital recording waveform or respiratory sound waveform) stored in the input data storage unit 102. .

  The evaluation unit 104 calculates an evaluation value (for example, an improvement degree) representing the wrinkle suction effect based on the count value of the stuttering counted by the count processing unit 103.

  The output unit 105 displays the count value of the stuttering counted by the count processing unit 103 and the index value calculated by the evaluation unit 104 on the display device.

  Next, a detailed functional configuration of the count processing unit 103 will be described with reference to FIG.

  The count processing unit 103 includes a high-pass filter 1031, a rectification unit 1032, a smoothing processing unit 1033, a rale candidate section detection unit 1034, a parameter extraction unit 1035, a bandpass filter 1036, and a stuttering determination unit 1037. And a count result storage unit 1038.

  The high-pass filter 1031 performs high-pass filter processing (for example, cut-off frequency fc = 50 Hz) on the digital recording waveform, and outputs the processing result to the rectifying unit 1032. The rectifying unit 1032 performs full-wave rectification on the processing result of the high-pass filter 1031 and outputs the processing result to the smoothing processing unit 1033. The smoothing processing unit 1033 performs a smoothing process on the processing result of the rectifying unit 1032 to detect a waveform envelope, and outputs the processing result to the ra sound candidate section detecting unit 1034.

  In the present embodiment, the respiratory sound waveform is not analyzed every frame (for example, 10 ms to 30 ms), but the respiration sound waveform is sampled with fineness of the sample level (for example, every 0.125 ms in the case of 8000 Hz sampling). By rectifying and smoothing the waveform component of the sound wave shape having a cutoff frequency fc or higher, a very short ra-tone candidate section, which is a characteristic of intermittent ra-cho, is detected.

  The ra sound candidate section detection unit 1034 detects a section in which the amplitude is equal to or greater than a predetermined threshold from the waveform that is the processing result of the smoothing processing section 1033, measures the duration time dur of the section, The processing result is output to the parameter extraction unit 1035.

  The band pass filter 1036 executes band pass filter processing (for example, a frequency band of 200 Hz to 1500 Hz) on the digital recording waveform, and outputs the processing result to the parameter extraction unit 1035.

  The parameter extraction unit 1035 performs spectrum analysis for each detected rale candidate section, extracts a characteristic frequency for the spectrum intensity (for example, the peak frequency fp or the center of gravity frequency fg for the spectrum intensity), and each rale candidate The duration time dur and the characteristic frequency (for example, fc or fg) are output to the stuttering discrimination unit 1037 for the section.

  The stuttering determination unit 1037 determines whether or not each rale candidate section is stuttering based on the criteria as shown in FIG. 1, counts the number of occurrences of stuttering, and stores it in the count result processing unit 1038. To do. Note that the number of occurrences of stuttering is calculated as the number of occurrences per unit time.

  In this way, it is possible to accurately identify and evaluate the roar among the rales.

  Next, processing contents of the information processing apparatus 100 according to the present embodiment will be described with reference to FIGS.

  For example, when a user such as a caregiver presses the measurement start button after pressing the pre-suction button, the input unit 101 records the respiratory sound before the sputum suction and stores the respiratory sound data in the input data storage unit 102 (FIG. 6: Step S1). Although an example in which a breathing sound is recorded until the measurement end button is pressed has been described, the recording may be automatically terminated when a predetermined time is measured.

  Then, for example, in response to an instruction from the input unit 101, the count processing unit 103 performs a stuttering counting process (step S3). Details of the stuttering counting process will be described with reference to FIGS.

  First, the high-pass filter 1031 performs high-pass filter processing with a cutoff frequency fc = 50 Hz on the digital recording waveform (respiratory sound waveform), and outputs the processing result to the rectifying unit 1032 (FIG. 7: Step S21).

  For example, considering a case where a respiratory sound waveform as shown in FIG. 8A is input, a low-frequency component that becomes noise in detecting stuttering such as heart sound is removed by high-pass filter processing, and FIG. A waveform as shown in b) is obtained.

  The rectifying unit 1032 performs full-wave rectification on the processing result of the high-pass filter process, and outputs the processing result to the smoothing processing unit 1033 (step S23). As described above, full-wave rectification is performed here in order to analyze the respiratory sound waveform after the high-pass filter at the sample level.

  For example, when full-wave rectification is performed on a waveform as shown in FIG. 8B, a waveform as shown in FIG. 8C is obtained. In this way, the amplitude of the waveform is converted to an absolute value.

  Further, the smoothing processing unit 1033 extracts the waveform envelope by smoothing the processing result of the full wave rectification using a Hanning window having a window length of, for example, about 15 ms, and the processing result is It outputs to the candidate area detection part 1034 (step S25).

  For example, by performing the smoothing process in this step on the waveform as shown in FIG. 8C, an envelope of the waveform can be obtained as shown in FIG. 8D.

  Then, the ra sound candidate section detection unit 1034 extracts a section having a predetermined threshold th1 or more as a ra sound candidate section from the processing result of the smoothing processing unit 1033, measures the continuation section length dur, and obtains the processing result. It outputs to the parameter extraction part 1035 (step S27).

  When a waveform envelope as shown in FIG. 8D is obtained, a peak portion is specified, and a section that is equal to or greater than a predetermined threshold th1 around the peak is shown in FIG. 8E. Is identified as a ra sound candidate section. The continuation section length dur may be measured by the parameter extraction unit 1035.

  Further, the band pass filter 1036 performs band pass filter processing for allowing only the 200 Hz to 1500 Hz band to pass through the digital recording waveform (respiratory sound waveform), and outputs the processing result to the parameter extraction unit 1035 (step S29). . This is because 200 Hz to 1500 Hz is a band in which the component of the ra sound appears.

  When bandpass filter processing is executed on a waveform as shown in FIG. 8A, a waveform as shown in FIG. 9B is obtained.

  And the parameter extraction part 1035 performs a spectrum analysis about each ra sound candidate area (step S31).

  In the case of FIG. 9B, only the waveform of the ra sound candidate section as shown in FIG. 9A is extracted, the spectrum analysis is executed, and the spectrum intensity from 200 Hz to 1500 Hz is calculated. Then, in this example, a spectrum analysis result as shown in FIG. 10 is obtained.

  Then, the parameter extraction unit 1035 identifies the characteristic frequency for each ra sound candidate section, and outputs the duration length dur and the characteristic frequency of each ra sound candidate section to the stuttering determination unit 1037 (step S33).

  In the example of FIG. 10, the peak frequency fp is 312 Hz, and the center-of-gravity frequency fg for the spectral intensity from 200 Hz to 1500 Hz is 361 Hz. Since the method using the peak frequency fp may be weak against noise or the like, the stability of parameter extraction can be increased by obtaining the centroid frequency fg. In the present embodiment, the peak frequency fp or the centroid frequency fg is specified.

  Then, the stuttering discrimination unit 1037 executes stuttering discrimination processing for each rale candidate section (step S35). Then, the process returns to the process of FIG. The stuttering discrimination process will be described in detail with reference to FIG.

  The stuttering determination unit 1037 identifies one unprocessed rale candidate section among one or more rale candidate sections to be processed (step S41). Then, the roaring determination unit 1037 determines whether or not the duration time dur of the identified rale candidate section is 25 ms or less (step S43). If it exceeds 25 ms, it is a continuous rarity or other condition. Therefore, when the condition of step S43 is not satisfied, the process proceeds to step S51. Although it may be specifically determined whether or not it is a continuous rarity according to FIG. 1, it is omitted in the present embodiment.

  On the other hand, if the duration time dur of the identified raji candidate section is 25 ms or less, the stuttering determination unit 1037 determines whether the duration time dur is 5 ms or less and the characteristic frequency is 500 to 1000 Hz. (Step S45). This is to determine whether or not it corresponds to a haircut sound.

  If the condition of step S45 is satisfied, the stuttering determination unit 1037 increments the stuttering count value by 1 (step S49). On the other hand, if the condition of step S45 is not satisfied, the stuttering determination unit 1037 determines whether the duration time dur is 10 ms or more and the characteristic frequency is 250 to 500 Hz (step S47). This is to determine whether or not it corresponds to a water bubble sound.

  If the condition of step S47 is not satisfied, the other state is indicated, and the process proceeds to step S51. On the other hand, if the condition of step S47 is satisfied, the stuttering determination unit 1037 increments the stuttering count by 1 (step S49).

  Then, the roaring determination unit 1037 determines whether or not an unprocessed rale candidate section exists (step S51). If there is an unprocessed rale candidate section, the process returns to step S41.

  On the other hand, if there is no unprocessed rale candidate section, the roaring determination unit 1037 calculates a count value per unit time (step S53). If the measurement time of breathing sound is a fixed time, this step can be omitted, but if the measurement time differs for each measurement, the count value per unit time is calculated from the time from the start of measurement to the end of measurement. calculate. The stuttering discrimination unit 1037 stores the count value per unit time in the count result processing unit 1038.

  In this way, stuttering can be extracted accurately and the number of occurrences can be accurately counted.

  Returning to the description of the processing in FIG. 6, when a user such as a caregiver presses the measurement start button after pressing the post-suction button, the input unit 101 records the respiratory sound after the sputum suction, and the respiratory sound data Is stored in the input data storage unit 102 (step S5).

  And the count process part 103 performs the count process of a stuttering with respect to the respiratory sound data obtained by step S5 (step S7). Although the processing target breathing sound data is different, the processing content is the same as the processing using FIGS.

  When step S7 is completed, the evaluation unit 104 calculates the evaluation value of the suction effect using the count value of the stuttering before and after the soot suction stored in the count result processing unit 1038 (step S9).

  For example, the evaluation value is calculated based on (the count value of the stuttering before the soot suction−the count value of the stuttering after the soot suction). Instead of a simple difference between values, a value obtained by dividing the difference by the count value of the stuttering sound before sucking may be adopted as the evaluation value. Furthermore, (the count value of the stuttering after the soot suction / the count value of the stuttering before the soot suction) may be used as the evaluation value. In this way, a numerical value representing the improvement degree is calculated.

  When the evaluation unit 104 outputs the evaluation value to the output unit 105, the output unit 105 displays on the display device the count value of the stuttering before the soot suction, the count value of the stuttering after the soot suction, and the evaluation value. (Step S11). Note that the count value of the stuttering sound before sucking may be displayed immediately after step S3.

  For example, a display as shown in FIG. 12 is performed. In this example, since the count value of the stuttering before the soot suction is “25” and the count value of the stuttering after the soot suction is “3”, (25−3) / 25 × 100 = 88% The evaluation value indicating the degree of improvement is displayed on the display device.

  In this way, even a caregiver or a patient's family with little experience can perform auscultation of stuttering and improve the quality of care at a facility or at home.

  Although the embodiment of the present invention has been described above, the present invention is not limited to this. For example, the functional block configuration shown in FIGS. 4 and 5 is an example, and may not match the program module configuration. Further, the processing flows shown in FIGS. 6, 7, and 11 are examples, and the processing order may be changed or the processing may be performed in parallel as long as the processing result does not change. For example, steps S21-S27 and step S29 can be executed in parallel.

  The display shown in FIG. 12 is an example, and the screen configuration can be changed variously.

  Furthermore, in the example described above, as shown in FIG. 2, the stuttering is counted immediately before and after the sputum suction. However, if this embodiment is applied, various evaluations are performed. Will be able to.

  For example, by calculating the difference between the current count value and the count value immediately after the previous soot suction, the number of increases in stuttering can be determined. Therefore, a second evaluation value based on such a difference may be defined, and the necessity for sputum suction may be converted into a numerical value based on the second evaluation value and presented. For example, an index value such as (current count value−count value immediately after the previous soot suction) / count value immediately after the previous soot suction may be employed. Alternatively, the current count value / the count value immediately after the previous sputum suction may be used.

  Similarly, by calculating the difference between the count value immediately before the previous sputum suction and the current count value, the degree of deterioration of the sputum state can be determined, so a third evaluation value based on such a difference is defined. In addition, the necessity for sputum suction may be converted into a numerical value by the third evaluation value and presented. For example, an index value such as (count value immediately before last soot suction−current count value) / count value just before the last soot suction may be employed. Alternatively, the current count value / the count value immediately before the previous sputum suction may be used.

  The information processing apparatus 100 described above is a computer apparatus, and is connected to a storage device such as a memory, a CPU (Central Processing Unit), a hard disk drive (HDD: Hard Disk Drive), and a flash memory, and a display device. A display control unit, a removable disk drive device, an input device, and a communication control unit for connecting to a network are connected by a bus. An operating system (OS) and an application program for executing the processing in this embodiment are stored in the storage device, and are read from the storage device to the memory when executed by the CPU. The CPU controls the display control unit, the communication control unit, and the drive device in accordance with the processing content of the application program, and performs a predetermined operation. Further, the data being processed is mainly stored in the memory, but may be stored in the storage device. In the embodiment of the present invention, an application program for performing the above-described processing is stored and distributed on a computer-readable removable disk, and installed from the drive device to the storage device. In some cases, the storage device is installed via a network such as the Internet and a communication control unit. Such a computer apparatus realizes various functions as described above by organically cooperating hardware such as the CPU and memory described above with programs such as the OS and application programs.

  The above-described embodiment can be summarized as follows.

  The information processing method according to the present embodiment includes (A) a step of counting a first number of occurrences of stuttering included in a breathing sound of a predetermined period recorded at a first timing, and (B) a second The step of counting the second number of occurrences of stuttering included in the breathing sound of a predetermined period recorded at the timing, and (C) the first timing and the second number of occurrences from the first number of occurrences and the second number of occurrences. And a calculation step for calculating an evaluation value representing a symptom change between the two timings.

  In this way, it becomes possible to easily recognize, for example, the effect of soot suction or the necessity of soot suction.

  In the information processing method described above, after (D) displaying the first button for instructing the count of the first occurrence and the second button for instructing the count of the second occurrence, The step of counting the first number of occurrences may be executed in response to a click on the first button, and the step of counting the second number of occurrences in response to a click on the second button may be executed. Then, after the calculation step described above, the evaluation value may be displayed. Such a user interface allows a caregiver or the like to easily operate.

  In the information processing method described above, the first timing may be before sputum suction and the second timing may be after sputum suction. In such a case, the evaluation value representing the symptom change may be an evaluation value representing the degree of symptom improvement due to sputum suction. A caregiver or the like can easily confirm whether the sputum suction has been effectively performed.

Further, in the information processing method described above, the first timing is before or after the first sputum suction, and the second timing is after the first sputum suction and the second sputum suction. Can be before. In such a case, the evaluation value indicating the symptom change may be an evaluation value indicating the necessity of sputum suction. A caregiver or the like can easily determine whether or not sputum suction is necessary.

  In addition, the step of counting the first number of occurrences or the number of second occurrences described above includes (1) executing a high-pass filter process on a breathing sound for a predetermined period; ) Rectifying the processing result of the high-pass filter process to extract a waveform envelope; (3) specifying a portion where the amplitude of the waveform envelope is a predetermined threshold or more as a candidate section; and (4) continuation of the candidate section. And determining whether the candidate section is a stuttering section based on the time length. In this way, it is possible to detect stuttering with high accuracy based on the duration time.

  Further, the step of counting the first number of occurrences or the second number of occurrences described above further includes the step of (5) performing a band pass filter process on the breathing sound for a predetermined period. You may make it. In this case, the step of determining whether or not the candidate section described above is a stuttering section is performed by (41) performing a spectrum analysis in the candidate section in the processing result of the bandpass filter process, Identifying a centroid frequency for the spectral intensity; and (42) determining whether the candidate section is a stuttering section based on the duration of the candidate section and the peak frequency or centroid frequency. You may make it. Since the characteristic frequency in the candidate section can also be specified, it is possible to detect stuttering with higher accuracy.

  A program for causing a computer to perform the above method can be created, and the program is a computer-readable storage medium or storage device such as a flexible disk, a CD-ROM, a magneto-optical disk, a semiconductor memory, or a hard disk. Stored in The intermediate processing result is temporarily stored in a storage device such as a main memory.

DESCRIPTION OF SYMBOLS 101 Input part 102 Input data storage part 103 Count processing part 104 Evaluation part 105 Output part 1031 High pass filter 1032 Rectification part 1033 Smoothing process part 1034 Ra sound candidate area detection part 1035 Parameter extraction part 1036 Band pass filter 1037 Stuttering discrimination part 1038 Count result storage

Claims (7)

Counting the first number of occurrences of stuttering included in breathing sounds of a predetermined period recorded at a first timing;
Counting a second number of occurrences of stuttering included in the breathing sound of the predetermined period recorded at a second timing;
A calculating step for calculating an evaluation value representing a symptom change between the first timing and the second timing from the first occurrence number and the second occurrence number;
To the computer ,
Counting the first number of occurrences or counting the second number of occurrences,
Performing a high-pass filter process on the breathing sound of the predetermined period;
Rectifying the processing result of the high-pass filter processing to extract a waveform envelope; and
Identifying a portion where the amplitude of the waveform envelope is a predetermined threshold or more as a candidate section;
Determining whether the candidate section is a section of stuttering based on the duration of the candidate section; and
If the candidate section is a stuttering section, incrementing the first number of occurrences or the second number of occurrences;
Including programs. After displaying a first button for instructing counting of the first occurrence count and a second button for instructing counting of the second occurrence count,
Performing the step of counting the first number of occurrences in response to a click on the first button;
Performing a step of counting the second number of occurrences in response to a click on the second button;
The program according to claim 1, wherein the evaluation value is displayed after the calculating step. The first timing is before sputum suction;
The second timing is after sputum suction;
The program according to claim 1 or 2, wherein the evaluation value representing the symptom change is an evaluation value representing a degree of symptom improvement by the sputum suction. The first timing is before or after the first sputum suction;
The second timing is after the first sputum suction and before the second sputum suction;
The program according to claim 1 or 2, wherein the evaluation value representing the symptom change is an evaluation value representing the necessity of sputum suction. Counting the first number of occurrences or counting the second number of occurrences,
Performing a band pass filter process on the breathing sound of the predetermined period ;
Identifying a centroid frequency for a peak frequency or spectrum intensity by performing a spectrum analysis within the candidate section in the processing result of the bandpass filter processing;
Further including
In the step of determining whether the candidate section is a section of stuttering ,

The program according to any one of claims 1 to 4, wherein the program further determines whether the candidate section is a stuttering section based on the peak frequency or the barycentric frequency . Counting the first number of occurrences of stuttering included in breathing sounds of a predetermined period recorded at a first timing;
Counting a second number of occurrences of stuttering included in the breathing sound of the predetermined period recorded at a second timing;
A calculating step for calculating an evaluation value representing a symptom change between the first timing and the second timing from the first occurrence number and the second occurrence number;
Including
Counting the first number of occurrences or counting the second number of occurrences,
Performing a high-pass filter process on the breathing sound of the predetermined period;
Rectifying the processing result of the high-pass filter processing to extract a waveform envelope; and
Identifying a portion where the amplitude of the waveform envelope is a predetermined threshold or more as a candidate section;
Determining whether the candidate section is a section of stuttering based on the duration of the candidate section; and
If the candidate section is a stuttering section, incrementing the first number of occurrences or the second number of occurrences;
An information processing method executed by a computer. Means for counting a first number of occurrences of stuttering included in a breathing sound of a predetermined period recorded at a first timing;
Means for counting a second number of occurrences of stuttering included in the breathing sound of the predetermined period recorded at a second timing;
Means for calculating an evaluation value representing a symptom change between the first timing and the second timing from the first occurrence count and the second occurrence count;
Have a,
Means for counting the first number of occurrences or means for counting the second number of occurrences;
A high-pass filter process is performed on the breathing sound of the predetermined period,
Rectifying the processing result of the high-pass filter processing to extract a waveform envelope,
Specify a portion where the amplitude of the waveform envelope is a predetermined threshold or more as a candidate section,
Based on the duration of the candidate section, determine whether the candidate section is a section of stuttering,
If the candidate section is a stuttering section, the first occurrence count or the second occurrence count is incremented.
Information processing device.

Images