TW201306794A - Method for extracting the feature of an abdominal breathing and a system using the same - Google Patents

Abstract

A method for extracting the feature of an abdominal breathing is disclosed, capable of extracting the feature of an abdominal breathing, without the requirement of a standard model of an abdominal breathing and the execution of a learning process being executed prior to the method for extracting the feature of an abdominal breathing. By means of computing a plurality of intrinsic mode functions corresponding to the abdominal breathing signal received, an Euler angle function and an instantaneous frequency function of each of the plurality of intrinsic mode functions, and comparing the plurality of instantaneous frequency function with a pre-determined zero-point threshold region, the method for extracting the feature of an abdominal breathing defines one of the plurality of instantaneous frequency function as an abdominal breathing feature function, which contains the feature of the abdominal breathing. In this way, the feature of an abdominal breathing is extracted.

Description

Abdominal breathing characteristic extraction method and abdominal breathing characteristic extraction system using the same

The present invention relates to a abdominal breathing feature extraction method and a abdominal breathing feature extraction system, in particular to a model that does not require the use of a abdominal breathing standard, and a subject does not need to receive a belly breathing feature beforehand. A abdominal breathing feature extraction method for learning how to perform a standard abdominal breathing learning procedure to extract the abdominal breathing characteristics of the subject, and a abdominal breathing feature extraction system.

In recent years, abdominal breathing has played a very important role in the rehabilitation and stress release procedures. However, a person who needs to perform abdominal breathing, such as a patient undergoing cardiac surgery, must be trained by a teacher (such as a doctor or nursing staff) for a period of time. However, the results of the aforementioned training, such as the correct rate of abdominal breathing performed by the patient, cannot be effectively evaluated because a standard model of abdominal breathing is still needed to mimic the patient's learning. In addition, the aforementioned abdominal breathing standard model does not meet the needs of various types of patients, such as the two abdominal breathing standard models for the elderly and children, respectively, which should be slightly different from each other.

As mentioned above, in order for the patient to learn the aforementioned abdominal breathing standard model, a learning program must be executed, and this learning procedure not only takes a considerable amount of time and a certain amount of money. More importantly, in this learning procedure, teachers who teach abdominal breathing and those who learn abdominal breathing must be in the same location, causing difficulties in transporting people, especially for older patients.

Therefore, the industry needs a model that does not require the use of a abdominal breathing standard, and a subject does not need to receive a learning procedure for learning a standard abdominal breathing before extracting their abdominal breathing characteristics. Abdominal respiratory feature extraction method for the abdominal breathing characteristics of the subject, and a abdominal respiratory feature extraction system.

The main object of the present invention is to provide a method for extracting abdominal breathing characteristics, which does not require the use of a standard model of abdominal breathing, and a subject does not need to receive a learning procedure before extracting the abdominal breathing characteristics. Next, the abdominal breathing characteristics of the subject were extracted.

Another object of the present invention is to provide a abdominal respiratory feature extraction system that does not require the use of a abdominal breathing standard model, and a subject does not need to receive a learning program before extracting its abdominal breathing characteristics. In this case, the abdominal breathing characteristics of the subject were extracted.

To achieve the above object, the abdominal respiratory feature extraction method of the present invention comprises the steps of: reading a abdominal breathing signal; performing an empirical modal analysis program on the abdominal breathing signal to calculate a plurality of intrinsic mode functions; Performing a Hilbert transform result based on the intrinsic mode functions and the intrinsic mode functions, respectively, to calculate that each of the intrinsic mode functions respectively has a Euler angle function; The pull angle function is differentially differentiated from time to calculate one of the instantaneous frequency functions of each of the intrinsic mode functions, and extracts a plurality of maximum values of each of the instantaneous frequency functions; and according to a specific order, each The maximum values of one of the instantaneous frequency functions are compared with a zero threshold range, and the instantaneous value is set when one of the instantaneous values of the instantaneous frequency functions is within the threshold of the zero point. The frequency function is a abdominal breathing characteristic function.

To achieve the above objective, the abdominal respiratory feature extraction system of the present invention comprises: a sensing module for sensing a abdominal breathing signal; and an arithmetic module coupled to the sensing module for extraction a belly breathing characteristic function; and a abdominal breathing feature output module coupled to the computing module to output the abdominal breathing characteristic function; wherein the computing module performs a abdominal breathing feature extraction In the method of the method, the abdominal breathing characteristic function is extracted from the abdominal breathing signal, and the abdominal breathing characteristic extraction method comprises the following steps: reading the abdominal breathing signal; performing an empirical mode on the abdominal breathing signal State analysis program for computing a plurality of intrinsic mode functions; calculating the result of each Hilbert transformation based on the intrinsic mode functions and the intrinsic mode functions, respectively, to calculate each of the intrinsic mode functions Each having a Euler angle function; these Euler angle functions are differentially differentiated from time to calculate an instantaneous frequency function of each of the intrinsic mode functions, Extracting a plurality of maxima of each of the instantaneous frequency functions; and comparing the maxima of each of the instantaneous frequency functions to a zero threshold range according to a particular order, and comparing one of the When these maxima of these instantaneous frequency functions are all within the threshold of this zero point, the instantaneous frequency function is set as a function of the abdominal breathing characteristic.

Since the abdominal breathing feature extraction method of the present invention can calculate a plurality of intrinsic mode functions corresponding to the abdominal breathing signals it receives, and successively calculate the Euler angle function of each intrinsic mode function. Subsequently, the abdominal breathing feature extraction method of the present invention can calculate one of the instantaneous frequency functions of each of the intrinsic mode functions by differentiating the Euler angle function of each of the intrinsic mode functions with respect to time. Then, the multiple maxima of each instantaneous frequency function are extracted and these maxima are compared to a zero threshold range. Finally, based on the results of the comparison, the abdominal breathing characteristic function of the received abdominal breathing signal can be set. Therefore, the abdominal respiratory feature extraction method of the present invention can directly extract the abdominal breathing characteristics of the subject from the received abdominal breathing signal without using a abdominal breathing standard model, and the subject does not need to extract. A learning procedure is accepted before the abdominal breathing feature.

In addition, a computing module for performing a abdominal breathing feature extraction method, a sensing module for sensing a abdominal breathing signal, and a abdominal breathing for outputting a abdominal breathing characteristic function are included. The characteristic output module, the abdominal respiratory feature extraction system of the present invention can extract the abdominal breathing characteristics of the subject without using a abdominal breathing standard model, and the subject does not need to extract the abdominal breathing characteristics before Accept a learning program in advance.

The remaining objects, advantages and innovations of the present invention will be apparent from the following detailed description and the accompanying drawings.

FIG. 1 is a schematic flow chart of a method for extracting abdominal breathing characteristics according to an embodiment of the present invention. Wherein, the abdominal respiratory feature extraction method of the present invention comprises the following steps:

(A) reading a belly breathing signal;

(B) performing an empirical modal analysis procedure on the abdominal breathing signal to calculate a plurality of intrinsic mode functions;

(C) calculating, according to the intrinsic mode function and the result of performing a Hilbert transform on the intrinsic mode functions, respectively, to calculate that each of the intrinsic mode functions respectively has a Euler angle function;

(D) differentiating the Euler angle functions from time to calculate an instantaneous frequency function of each of the intrinsic mode functions and extracting a plurality of maxima of each of the instantaneous frequency functions;

(E) comparing the maxima of each of the instantaneous frequency functions with a zero threshold range according to a particular order, and when comparing one of the magnitudes of the instantaneous frequency functions, the maxima are located When the zero threshold is within the range, the instantaneous frequency function is set to a belly breathing characteristic function.

Hereinafter, the detailed flow of each step included in the abdominal breathing feature extraction method according to an embodiment of the present invention will be described in detail with reference to the drawings.

First, please refer to FIG. 2, which is a schematic diagram of a abdominal respiratory feature extraction system according to an embodiment of the present invention. A subject wears two fixed straps 21, 22 on their chest and abdomen to extract their abdominal breathing characteristics. Next, the subject performs abdominal breathing or chest breathing separately until it receives further instructions.

When the abdominal breathing feature extraction method of one embodiment of the present invention is performed, the chest displacement and the abdominal displacement of the subject are extracted separately and simultaneously by the two fixed straps 21, 22. The two fixed straps 21, 22 convert these displacements (thorax displacement and abdominal displacement) into corresponding electronic signals by using a piezoelectric element. Once these electronic signals (which will be referred to as breathing signals) are derived, these electronic signals are transmitted to a computer and/or portable device 23 by various possible signal transmission methods, such as wired or wireless. The computer and/or portable device 23 stores a computer program in its memory unit to perform the abdominal breathing feature extraction method according to an embodiment of the present invention.

When the computer and/or portable device 23 of the abdominal breathing characteristic extraction receives the abdominal breathing signal, that is, the step (A) of the abdominal breathing characteristic extraction method according to an embodiment of the present invention, the abdominal type according to an embodiment of the present invention Step (B) of the respiratory feature extraction method is carried out successively. As shown in FIG. 1, in step (B), an empirical mode decomposition process is performed on the abdominal breathing signal described above to calculate the plural of each abdominal breathing signal. An intrinsic mode function (IMF).

In the present embodiment, when the subject performs abdominal breathing or chest breathing, a total of two breathing signals are extracted through the two fixed straps 21, 22. These breathing signals are then processed by an empirical modal analysis program (referred to as the EMD program) to calculate a plurality of intrinsic mode functions for each of the abdominal breathing signals. In this embodiment, a plurality of intrinsic mode functions respectively possessed by the two breathing signals are calculated.

For example, as shown in FIG. 2, after the respiratory signal extracted by the fixed strap 21 is received by the computer and/or the portable device 23, the computer and/or the portable device 23 will breathe the signal (as shown in FIG. 3). The chest breathing signal shown) performs the aforementioned empirical modal analysis procedure to obtain a plurality of intrinsic mode functions corresponding to the respiratory signal (thoracic breathing signal), the so-called IMF. In the example shown in FIG. 3A, the chest breathing signal shown in FIG. 3A can deconstruct 16 intrinsic mode functions. However, it should be noted that the number of intrinsic mode functions that each respiratory signal can deconstruct is not limited to this, and may be 12 or 20 depending on the intrinsic characteristic of the respiratory signal.

Once the 16 intrinsic mode functions of the chest breathing signal shown in Figure 3 have been derived, the eleventh intrinsic mode function is shown in Figure 3B as an example. It should be noted that the mode function is randomly selected and displayed in FIG. 3B, and this selection (the eleventh intrinsic mode function) should not be used to limit the scope of the present invention.

The following eight patterns, from FIG. 3A to FIG. 6B, are used to display four breathing signals (FIG. 3A, FIG. 4A, FIG. 5A, and FIG. 6A), and after performing an empirical modal analysis program, One of the intrinsic mode functions selected from the plurality of intrinsic mode functions. As shown in the foregoing, Fig. 3A shows a case where the chest breathing signal extracted by the fixed band 21 placed on the chest of the subject changes with time when a subject performs chest breathing. After performing the empirical modal analysis program, one of the 16 intrinsic mode functions (the 11th intrinsic mode function) selected from the 16 intrinsic mode functions obtained by the operation is displayed in FIG. 3B.

According to the same mode, Fig. 4A shows a case where the abdominal breathing signal extracted by the fixed band 22 placed on the abdomen of the subject changes with time when a subject performs chest breathing. After performing the empirical modal analysis program, one of the intrinsic mode functions selected from the plurality of intrinsic mode functions obtained by the operation is displayed in FIG. 4B. Further, Fig. 5A shows a case where the chest breathing signal extracted by the fixed band 21 placed on the chest of the subject changes with time when a subject performs abdominal breathing. After performing the empirical modal analysis program, one of the intrinsic mode functions selected from the plurality of intrinsic mode functions obtained by the operation is displayed in FIG. 5B.

On the other hand, Fig. 6A shows a case where the abdominal breathing signal extracted by the fixed band 22 placed on the abdomen of the subject changes with time when a subject performs abdominal breathing. After performing the empirical modal analysis program, one of the intrinsic mode functions selected from the plurality of intrinsic mode functions obtained by the operation is shown in FIG. 6B.

The detailed steps of performing an empirical modal analysis procedure for a signal (such as the respiratory signal of the present invention) are well known in the art, for example in the field of nonlinear and non-steady state data processing. For this reason, a detailed description of the execution steps of the empirical modal analysis program will not be repeated here.

After a plurality of intrinsic mode functions of the abdominal breathing signal (such as the abdominal breathing signal shown in FIG. 6A) are calculated, the step (C) of the abdominal breathing feature extraction method according to an embodiment of the present invention is carried out. In this embodiment, after the abdominal breathing signal shown in FIG. 6A is operated, a total of 16 intrinsic mode functions can be deconstructed. However, after the execution of the aforementioned empirical modal analysis program, the number of intrinsic mode functions that can be deconstructed (operated) is not limited thereto, and is deconstructed according to the essential characteristics of the abdominal breathing signal. The number of intrinsic mode functions can be an integer between 1 and 16. Further, in these intrinsic mode functions, each of the intrinsic mode functions respectively has a characteristic frequency, and the values of each of the feature frequencies are different from each other.

As shown in FIG. 1, in step (C), each of the intrinsic mode functions respectively has an Euler angle function according to the intrinsic mode function and performing a Hilbert on the intrinsic mode function respectively. The result obtained after the Hilbert transform is calculated. In this embodiment, since there are 16 intrinsic mode functions, 16 results can be obtained after performing Hilbert conversion. Therefore, after performing step (C), 16 Euler angle functions can be obtained.

Due to the detailed steps of performing a Hilbert transform for a function, it is well known in the industry, for example in the field of signal processing. For this reason, a detailed description of the execution steps of the Hilbert conversion will not be repeated here.

After performing a Hilbert transform on the 16 intrinsic mode functions, 16 Euler angle functions are obtained. In the present embodiment, the Euler angle function is a function for describing a relationship between Euler angle and time in an Euler formula, and this Euler type can be expressed. for:

Re ^{j θ} = r (cos θ + j sin θ) (1)

Where r is the magnitude and θ is the Euler angle.

Therefore, since the Euler angle function is calculated based on an intrinsic mode function and a result obtained by performing a Hilbert transformation on the intrinsic mode function, the Euler angle function can be expressed as:

Wherein, r _a is the a-th line intrinsic mode function (a ^th intrinsic mode function) intensity, θ _a line is a th Euler angles intrinsic mode function, H represents the Hilbert conversion.

As described above, after performing step (C), 16 Euler angle functions can be calculated. Next, in the step (D) of the abdominal breathing feature extraction method according to an embodiment of the present invention, each intrinsic mode is calculated by differentiating the Euler angle function of each intrinsic mode function from time to time. The function has one of the instantaneous frequency functions. And this instantaneous frequency function can be expressed as:

Where ω _a is the instantaneous frequency function of the a-a intrinsic mode function, and θ ^a is the Euler angle of the a-a intrinsic mode function.

As mentioned above, since there are a total of 16 intrinsic mode functions corresponding to the abdominal breathing signal shown in Fig. 6A (this abdominal breathing signal is placed on the abdomen of the subject when the subject performs abdominal breathing) The fixed band 22 is extracted, so after the time-biasing of the Euler angle functions of the 16 intrinsic mode functions, 16 instantaneous frequency functions can be calculated. Among the 16 instantaneous frequency functions, the case where 5 instantaneous frequency functions change with time is shown in FIGS. 7A to 7E. It should be noted that these five are selected and shown in Figures 7A through 7E. The five instantaneous frequency functions are chosen arbitrarily, and this selection should not be used to limit the scope of the invention.

Once all of the instantaneous frequency functions have been computed, each of the instantaneous frequency functions (in this embodiment, 16 instantaneous frequency functions) has a plurality of maximum values that are extracted. Further, as shown in these five patterns (Figs. 7A to 7E), the respective maximum values of the five instantaneous frequency functions are a plurality of local maximum values.

Finally, in the step (E) of the abdominal breathing feature extraction method according to an embodiment of the present invention, the plurality of maximum values of the instantaneous frequency function of each intrinsic mode function are respectively associated with a zero point threshold according to a specific sequence. Zero-point threshold region comparison. In the present embodiment, this particular order refers to an order in which an intrinsic mode function having a higher characteristic frequency is arranged to another intrinsic mode function having a lower characteristic frequency. In other words, these instantaneous frequency functions are sequentially arranged according to the values of the characteristic frequencies of the intrinsic mode functions to which they correspond.

Therefore, in this embodiment, a plurality of maximum values of an instantaneous frequency function corresponding to an intrinsic mode function having the highest characteristic frequency are first compared with the aforementioned zero point threshold range, for example, as shown in FIG. 7A. Instantaneous frequency function.

As shown in FIG. 7A, the zero threshold value range 71 refers to a range in which the instantaneous frequency function has a value between -0.5 and 0.5, so the zero threshold value range 71 is also referred to as a zero point threshold value of ±0.5. As can be easily seen from FIG. 7A, the plurality of maximum values of the instantaneous frequency function are mostly above the upper limit of the zero threshold range 71 (top-limit). Therefore, it is not possible to obtain from the instantaneous frequency function shown in FIG. 7A the result that a plurality of maximum values of an instantaneous frequency function are within the zero threshold value range 71.

Next, the plurality of maxima of the instantaneous frequency function (the instantaneous frequency function as shown in FIG. 7B) possessed by the next intrinsic mode function in the specific sequence described above are compared with the zero threshold range 71, respectively. As can be readily seen from Figure 7B, the complex maximum frequency portion of the instantaneous frequency function is above the upper limit of the zero threshold range 71. Therefore, it is also impossible to obtain from the instantaneous frequency function shown in Fig. 7B the result that a plurality of maximum values of an instantaneous frequency function are within the zero threshold value range 71.

Then, in the same manner, the plurality of maxima of the instantaneous frequency function (the instantaneous frequency function shown in FIG. 7C) of the next intrinsic mode function in the specific order described above are respectively compared with the zero threshold range 71. . As can be readily seen from Figure 7C, the maximum of the plurality of maximum values of the instantaneous frequency function, i.e., the global maximum value, is above the upper limit of the zero threshold range 71. Therefore, it is also impossible to obtain from the instantaneous frequency function shown in Fig. 7C the result that a plurality of maximum values of an instantaneous frequency function are within the zero threshold value range 71.

Next, in the same manner, the plurality of maxima of the instantaneous frequency function (the instantaneous frequency function shown in FIG. 7D) of the next intrinsic mode function in the specific sequence described above and the zero threshold range 71 respectively. Comparison. As can be readily seen from Figure 7D, the maximum of the plurality of maximum values of the instantaneous frequency function, i.e., the wide-area maximum at around 93 seconds, is above the upper limit of the zero threshold range 71. Therefore, it is also impossible to obtain from the instantaneous frequency function shown in Fig. 7D the result that a plurality of maximum values of an instantaneous frequency function are within the zero threshold value range 71.

However, in the instantaneous frequency function (as shown in the instantaneous frequency function shown in FIG. 7E) of the next intrinsic mode function in the specific order described above, the maximum of the plurality of maxima is below the zero threshold range 71. The upper limit, and the minimum of the plurality of maxima is above the lower limit of the zero threshold range 71. Therefore, from the instantaneous frequency function shown in FIG. 7E, a result that a plurality of maximum values of an instantaneous frequency function are within the zero threshold value range 71 can be obtained.

At this time, when a result that an instantaneous frequency function has a plurality of maximum values within the zero threshold value range 71, the instantaneous frequency function having a plurality of maximum values is set to a abdominal breathing characteristic function (abdominal breathing). Feature function). In the present embodiment, the instantaneous frequency function shown in Fig. 7E is set to a belly breathing characteristic function corresponding to the abdominal breathing signal shown in Fig. 6. In addition to this, in the present embodiment, the frequency at which a plurality of minimum values of the abdominal breathing characteristic function occur is the frequency at which the subject performs abdominal breathing.

It should be noted that the abdominal breathing characteristic extraction method according to an embodiment of the present invention corresponds to the remaining abdominal breathing signals (the abdominal breathing signal shown in FIG. 5A and the abdominal breathing signal shown in FIG. 6A). The abdominal breathing characteristic function can be derived.

Please refer to FIG. 8 , which is a schematic diagram of a abdominal respiratory feature extraction system according to another embodiment of the present invention. As shown in the figure, a belly breathing feature extraction system according to another embodiment of the present invention includes a sensing module 81, a computing module 82 coupled to the sensing module 81, and a belly coupled to the computing module 82. Breathing feature output module 83. The sensing module 81 is used to sense a abdominal breathing signal, and the computing module 82 is used to extract a abdominal breathing characteristic function. In addition, the abdominal breathing feature output module 83 is for outputting the abdominal breathing characteristic function.

On the other hand, the arithmetic module 82 extracts the abdominal breathing characteristic function from the abdominal breathing signal by performing a abdominal breathing feature extraction method. As shown in FIG. 9, the abdominal breathing feature extraction method performed by the operation module of the abdominal respiratory feature extraction system according to another embodiment of the present invention includes the following steps:

(A) reading a belly breathing signal;

(B) performing an empirical modal analysis procedure on the abdominal breathing signal to calculate a plurality of intrinsic mode functions;

(C) calculating, according to the intrinsic mode function and the result of performing a Hilbert transform on the intrinsic mode functions, respectively, to calculate that each of the intrinsic mode functions respectively has a Euler angle function;

(D) differentiating the Euler angle functions from time to calculate an instantaneous frequency function of each of the intrinsic mode functions and extracting a plurality of maxima of each of the instantaneous frequency functions;

(E) comparing the maxima of each of the instantaneous frequency functions with a zero threshold range according to a particular order, and when comparing one of the magnitudes of the instantaneous frequency functions, the maxima are located When the zero threshold is within the range, the instantaneous frequency function is set to a belly breathing characteristic function.

In the present embodiment, the sensing module 81 of the abdominal respiratory feature extraction system of another embodiment of the present invention is integrated onto a fixed strap. For example, a fixed strap that allows a subject to wear on their body, especially for the subject to wear on their abdomen. In addition, as shown in FIG. 8, the sensing module 81 includes a piezoelectric unit 811 for extracting the body displacement of the subject, and an analog digital conversion unit 812 electrically connected to the piezoelectric unit 811 to The analog signal from the piezoelectric unit 811 is converted into a corresponding digital signal. In addition, in order to reduce the size and weight of the sensing module 81, the sensing module 81 is a system on chip (SoC).

In addition, as shown in FIG. 8, the abdominal breathing feature output module 83 includes a wireless transmission unit 831 for transmitting the abdominal breathing characteristic function to a remote end through wireless transmission, such as 3G or WiFi. Server (not shown).

In the abdominal breathing feature extraction system according to another embodiment of the present invention, the abdominal breathing feature extraction method performed by the computing module 82 is the same as the abdominal breathing feature extraction method according to an embodiment of the present invention, so the present invention A detailed description of each step of the abdominal breathing feature extraction method performed by the arithmetic module 82 of the abdominal respiratory feature extraction system of an embodiment will not be repeated herein to simplify the description of the present embodiment.

The above-mentioned embodiments are merely examples for convenience of description, and the scope of the claims is intended to be limited to the above embodiments.

21, 22. . . Fixed strap

twenty three. . . Computer and / or portable device

81. . . Sensing module

82. . . Computing module

83. . . Abdominal breathing feature output module

811. . . Piezoelectric unit

812. . . Analog digital conversion unit

831. . . Wireless transmission unit

1 is a schematic flow chart of a method for extracting abdominal breathing characteristics according to an embodiment of the present invention.

2 is a schematic diagram of a abdominal respiratory feature extraction system in accordance with an embodiment of the present invention.

Fig. 3A shows a situation in which a chest breathing signal extracted by a fixed band placed on the chest of a subject changes over time when a subject performs chest breathing.

Figure 3B shows the situation in which an intrinsic mode function changes over time in a plurality of intrinsic mode functions corresponding to the chest breathing signal shown in Figure 3A.

Fig. 4A shows a case where a abdominal breathing signal extracted by a fixed band placed on the abdomen of a subject changes with time when a subject performs chest breathing.

Figure 4B shows the situation in which an intrinsic mode function changes over time in a plurality of intrinsic mode functions corresponding to the abdominal breathing signals shown in Figure 4A.

Fig. 5A shows a situation in which a chest breathing signal extracted by a fixed band placed on the chest of a subject changes with time when a subject performs abdominal breathing.

Figure 5B shows the situation in which an intrinsic mode function changes over time in a plurality of intrinsic mode functions corresponding to the chest breathing signal shown in Figure 5A.

Fig. 6A shows a case where a abdominal breathing signal extracted by a fixed band placed on the abdomen of a subject changes with time when a subject performs abdominal breathing.

Fig. 6B shows the case where a certain intrinsic mode function changes over time in a plurality of intrinsic mode functions corresponding to the abdominal breathing signals shown in Fig. 6A.

7A to 7E are diagrams showing a plurality of intrinsic mode functions corresponding to the abdominal breathing signals shown in Fig. 6A, wherein the five intrinsic mode functions respectively have one of the instantaneous frequency functions varying with time.

Figure 8 is a schematic illustration of a abdominal respiratory feature extraction system in accordance with another embodiment of the present invention.

9 is a schematic flow chart of a belly breathing feature extraction method performed by a computing module of a belly breathing feature extraction system according to another embodiment of the present invention.

(The figure is a flow chart, so there is no component symbol)

Claims (16)

A method for extracting abdominal breathing characteristics includes the steps of: reading a abdominal breathing signal; performing an empirical modal analysis program on the abdominal breathing signal to calculate a plurality of intrinsic mode functions; according to the intrinsic mode functions And respectively performing a Hilbert transform on the intrinsic mode functions to calculate that each of the intrinsic mode functions respectively has a Euler angle function; and the Euler angle functions are time-biased Deriving to calculate an instantaneous frequency function of each of the intrinsic mode functions and extracting a plurality of maxima of each of the instantaneous frequency functions; and each of the instantaneous frequency functions according to a particular order The maximum value is compared with a zero threshold range, and when one of the maximum values of the instantaneous frequency functions is compared within the zero threshold range, the instantaneous frequency function is set to a belly type Breathing characteristic function. The abdominal breathing characteristic extraction method according to claim 1, wherein the Euler angle function is a Euler type, a Euler angle as a function of time, and the Euler type is re ^{j θ} = r (cos θ + j sin θ) where r is the intensity and θ is the Euler angle. The abdominal breathing feature extraction method of claim 1, wherein the maximum value of each of the instantaneous frequency functions is a local maximum. The abdominal breathing feature extraction method of claim 1, wherein the number of the intrinsic mode functions is between 1 and 16, and each of the intrinsic mode functions respectively has a characteristic frequency of a different value. The method according to claim 1, wherein the specific sequence refers to an intrinsic mode function having a higher characteristic frequency, and is arranged to another intrinsic mode function having a lower characteristic frequency. order. The abdominal breathing characteristic extraction method according to claim 1, wherein the zero threshold value range refers to a range of a function of the instantaneous frequency between -0.5 and 0.5. The abdominal breathing characteristic extraction method according to claim 1, wherein in the abdominal breathing characteristic function, the frequency at which the minimum value occurs is a frequency at which a subject performs abdominal breathing. A abdominal breathing feature extraction system includes: a sensing module for sensing a abdominal breathing signal; and an arithmetic module coupled to the sensing module to extract a abdominal breathing a feature function; and a belly breathing feature output module coupled to the computing module to output the abdominal breathing feature function; wherein the computing module is performed by performing a abdominal breathing feature extraction method Extracting the abdominal breathing characteristic function from the abdominal breathing signal, and the abdominal breathing characteristic extraction method comprises the steps of: reading the abdominal breathing signal; performing an empirical modal analysis program on the abdominal breathing signal Computation of a plurality of intrinsic mode functions; respectively, performing a Hilbert transform on the intrinsic mode functions and the intrinsic mode functions, respectively, to calculate each of the intrinsic mode functions respectively a Euler angle function; the Euler angle functions are differentiated from time to calculate an instantaneous frequency function of each of the intrinsic mode functions, and each of the instants is extracted a plurality of maxima of the frequency function; and comparing the maxima of each of the instantaneous frequency functions to a range of zero threshold values in accordance with a particular order, and comparing one of the instantaneous frequency functions When the equal magnitude is within the threshold value of the zero point, the instantaneous frequency function is set to be the abdominal breathing characteristic function. The abdominal breathing characteristic extraction system of claim 8, wherein the sensing module is disposed on a fixed strap, and the sensing module comprises a piezoelectric unit and an analog digital conversion unit. . The abdominal respiratory feature extraction system of claim 8, wherein the sensing module is a system single wafer. The abdominal breathing feature extraction system of claim 8, wherein the abdominal breathing feature output module comprises a wireless transmission unit. The abdominal breathing characteristic extraction system according to claim 8, wherein the Euler angle function is a Euler type, a Euler angle as a function of time, and the Euler type is re ^{j θ} = r (cos θ + j sin θ) where r is the intensity and θ is the Euler angle. The abdominal respiratory feature extraction system of claim 8, wherein the number of the intrinsic mode functions is between 1 and 16, and each of the intrinsic mode functions respectively has a characteristic frequency of a different value. The abdominal respiratory feature extraction system of claim 8, wherein the specific sequence refers to an intrinsic mode function having a higher characteristic frequency, and is arranged to another intrinsic mode function having a lower characteristic frequency. order. The abdominal respiratory characteristic extraction system of claim 8, wherein the zero threshold value range refers to a range of instantaneous frequency functions ranging between -0.5 and 0.5. The abdominal breathing characteristic extraction system according to claim 8, wherein in the abdominal breathing characteristic function, the frequency at which the minimum values occur is a frequency at which a subject performs abdominal breathing.