System and method for breathing pattern extraction from PPG signals

1. A processor-implemented method for extracting breathing patterns from photoplethysmogram (PPG) signals, the method comprising:

obtaining, from an accelerometer sensor embodied in a wearable device, an accelerometer signal indicative of motion of a subject;

pre-processing, via one or more hardware processors, raw PPG signals to obtain the PPG signals associated with a window size and a sampling frequency suitable for the breathing pattern extraction, wherein the step of pre-processing the raw PPG signals includes interpolating signals received from the accelerometer sensor and to bring the PPG signals to a uniform sampling frequency;

computing, via the one or more hardware processors, mean and standard deviation values of the accelerometer signal to determine whether the subject is in motion; and

performing, via the one or more hardware processors, based on the determination, one of:

discarding the PPG signals captured from the wearable device upon determination of the subject to be in motion, and

generating a plurality of filter coefficients upon determination of the subject to be in rest position;

designing, via the one or more hardware processors, a filter for extracting the breathing patterns from the PPG signals, wherein designing the filter comprises:

defining filter specifications for extraction of a breathing pattern from the PPG signals, wherein the filter specifications comprises a type, an order, and a cut-off frequency of the filter, and wherein the filter specifications for the extraction of the breathing pattern further comprises Infinite Impulse response (IIR) Butter-worth band-pass filter of order 4, with the cut-off frequency of 0.2-1 Hz to include a slow breathing rate of 12-15 Breaths per Minute (BrPM) up to fast breathing rate of 50-55 BrPM,

generating a transfer function associated with the filter specifications, and

computing the plurality of filter coefficients using filtfilt function for allowing filtering of the PPG signals, wherein the filtfilt function performs zero-phase digital filtering by processing an input data in both the forward and reverse directions;

generating, using the filter comprising the plurality of filter coefficients, filtered PPG signals by removing DC component from the PPG signals obtained from the wearable device configured to be worn by the subject via the one or more hardware processors, the filtered PPG signals indicative of the breathing pattern of the subject;

normalizing, via the one or more hardware processors, the filtered PPG signals with respect to the mean and standard deviation values;

analyzing, via the one or more hardware processors, the filtered PPG signal in frequency domain, by computing Fourier Transform (FFT) and obtaining power spectrum, wherein from the power spectrum, a highest isolated peak is taken as breathing frequency and is used to calculate the breathing rate in BrPM;

detecting, via the one or more hardware processors, anomalous behavior in stress-bearing capability of body of the subject based on the calculated breathing rate, identifying signs of cardiopulmonary disorders at an earliest stage, in the subject based on the detection of anomalous behavior, wherein the cardiopulmonary disorders include fatigue during physically intensive spells, palpitation and dyspnea; and

providing, via the one or more hardware processors, an opportunistic sensing of one or more physiological parameters based on the detected anomalous behavior, wherein the one or more physiological parameters are recorded before and after an activity.

2. The method as claimed in claim 1 , wherein the window size is around 10 seconds.

3. A system for extracting breathing patterns from photoplethysmogram (PPG) signals, the system comprising:

one or more memories; and

one or more hardware processors, the one or more memories coupled to the one or more hardware processors, wherein the one or more hardware processors are capable of executing programmed instructions stored in the one or more memories to:

obtain, from an accelerometer sensor embodied in a wearable device, an accelerometer signal indicative of motion of a subject;

pre-process, raw PPG signals to obtain the PPG signals associated with a window size and a sampling frequency suitable for the breathing pattern extraction, wherein pre-processing the raw PPG signals includes interpolating signals received from the accelerometer sensor and to bring the PPG signals to a uniform sampling frequency;

compute, mean and standard deviation values of the accelerometer signal to determine whether the subject is in motion; and

perform, based on the determination, one of:

discard the PPG signals captured from the wearable device upon determination of the subject to be in motion, and

generate the plurality of filter coefficients upon determination of the subject to be in rest position;

design a filter for extracting the breathing patterns from the PPG signals, wherein designing the filter comprises:

defining filter specifications for extraction of a breathing pattern from the PPG signals, wherein the filter specifications comprising a type, an order and a cut-off frequency of the filter, and wherein the filter specifications for the extraction of the breathing pattern further comprises Infinite Impulse response (IIR) Butter-worth band-pass filter of order 4, with the cut-off frequency of 0.2-1 Hz to include a slow breathing rate of 12-15 Breaths per Minute (BrPM) up to fast breathing rate of 50-55 BrPM,

generating a transfer function associated with the filter specifications, and

computing the plurality of filter coefficients using filtfilt function for allowing filtering of the PPG signals, wherein the filtfilt function performs zero-phase digital filtering by processing an input data in both the forward and reverse directions;

generate, using the filter comprising the plurality of filter coefficients, filtered PPG signals by removing DC component from the PPG signals obtained from the wearable device configured to be worn by the subject, the filtered PPG signals indicative of the breathing pattern of the subject;

normalize, the filtered PPG signals with respect to the mean and standard deviation values;

analyze the filtered PPG signal in frequency domain, by computing Fourier Transform (FFT) and obtaining power spectrum, wherein from the power spectrum, a highest isolated peak is taken as breathing frequency and is used to calculate the breathing rate in BrPM; and

detect anomalous behavior in stress-bearing capability of body of the subject based on the calculated breathing rate, to identify signs of cardiopulmonary disorders at an earliest stage, in the subject based on the detection of anomalous behavior, wherein the cardiopulmonary disorders include fatigue during physically intensive spells, palpitation and dyspnea; and

provide an opportunistic sensing of one or more physiological parameters based on the detected anomalous behavior, wherein the one or more physiological parameters are recorded before and after an activity.

4. The system as claimed in claim 3 , wherein the window size is around 10 seconds.

5. One or more non-transitory machine readable information storage mediums comprising one or more instructions which when executed by one or more hardware processors causes the one or more hardware processor to perform a method for extracting breathing patterns from photoplethysmogram (PPG) signals, said method comprising:

obtaining, from an accelerometer sensor embodied in a wearable device, an accelerometer signal indicative of motion of a subject;

pre-processing raw PPG signals to obtain the PPG signals associated with a window size and a sampling frequency suitable for the breathing pattern extraction, wherein the step of pre-processing the raw PPG signals includes interpolating signals received from the accelerometer sensor and to bring the PPG signals to a uniform sampling frequency;

computing, mean and standard deviation values of the accelerometer signal to determine whether the subject is in motion; and

performing, based on the determination, one of:

discarding the PPG signals captured from the wearable device upon determination of the subject to be in motion, and

generating the plurality of filter coefficients upon determination of the subject to be in rest position;

designing, via the one or more hardware processors, a filter for extracting the breathing patterns from the PPG signals, wherein designing the filter comprises:

defining filter specifications for extraction of a breathing pattern from the PPG signals, wherein the filter specifications comprises a type, an order, and a cut-off frequency of the filter, and wherein the filter specifications for the extraction of the breathing pattern further comprises Infinite Impulse response (IIR) Butter-worth band-pass filter of order 4, with the cut-off frequency of 0.2-1 Hz to include a slow breathing rate of 12-15 Breaths per Minute (BrPM) up to fast breathing rate of 50-55 BrPM,

generating a transfer function associated with the filter specifications, and

computing the plurality of filter coefficients using filtfilt function for allowing filtering of the PPG signals, wherein the filtfilt function performs zero-phase digital filtering by processing an input data in both the forward and reverse directions;

generating, using the filter comprising the plurality of filter coefficients, filtered PPG signals by removing DC component from the PPG signals obtained from the wearable device configured to be worn by the subject via the one or more hardware processors, the filtered PPG signals indicative of the breathing pattern of the subject;

normalizing the filtered PPG signals with respect to the mean and standard deviation values;

analyzing, the filtered PPG signal in frequency domain, by computing Fourier Transform (FFT) and obtaining power spectrum, wherein from the power spectrum, a highest isolated peak is taken as breathing frequency and is used to calculate the breathing rate in BrPM; and

detecting, anomalous behavior in stress-bearing capability of body of the subject based on the calculated breathing rate, identify signs of cardiopulmonary disorders at an earliest stage, in the subject based on the detection of anomalous behavior, wherein the cardiopulmonary disorders include fatigue during physically intensive spells, palpitation and dyspnea; and

providing an opportunistic sensing of one or more physiolo ig cal parameters based on the detected anomalous behavior, wherein the one or more physiological parameters are recorded before and after an activity.

6. The one or more non-transitory machine readable information storage mediums of claim 5 , wherein the window size is around 10 seconds.