Drift, noise, and motion artifact correction method for photoplethysmogram (PPG) signals and system thereof

1. A method for processing a photoplethysmogram (PPG) signal, the method comprising:

removing, by a device, a baseline drift from the PPG signal;

obtaining, by the device, a drift removed signal, based on removing the drift from the PPG signal;

filtering, by the device, the drift removed signal;

obtaining, by the device, a filtered signal, based on filtering the drift removed signal;

performing, by the device, motion artifact correction on the filtered signal, wherein the performing the motion artifact correction comprises:

obtaining a plurality of wavelet coefficients by applying a continuous wavelet transform (CWT) to the filtered signal;

identifying motion artifact corrupted coefficients from the plurality of wavelet coefficients, as outliers, based on an energy percentage of the plurality of wavelet coefficients being greater than a threshold for a same translation parameter and more than two scaling parameters; and

updating the motion artifact corrupted coefficients by reducing an intensity of the identified motion artifact corrupted coefficients; and

obtaining, by the device, a corrected signal, based on the updated motion artifact corrupted coefficients.

2. The method of claim 1 , further comprising:

measuring at least one of a blood pressure, a heart rate, an oxygen saturation, a glucose level, and a mental stress level based on the corrected signal.

3. The method of claim 1 , wherein removing the baseline drift from the PPG signal comprises:

performing a multi-level decomposition on the PPG signal using discrete wavelet transform (DWT);

obtaining approximation coefficients, based on performing the multi-level decomposition on the PPG signal;

obtaining a drift signal from the approximation coefficients at a pre-defined level;

removing the drift signal from the PPG signal; and

obtaining the drift removed signal based on removing the drift signal from the PPG signal.

4. The method of claim 1 , wherein the plurality of wavelet coefficients comprise approximated coefficients and detailed coefficients, and

wherein the approximated coefficients are obtained by using an inner product of the filtered signal of the PPG signal and a father wavelet, and the detailed coefficients are obtained by using an inner product of the filter signal of the PPG signal and a mother wavelet.

5. The method of claim 1 , wherein filtering the drift removed signal is performed by using a filter of order n, and

wherein the order of the filter n is 2, 3, 5, or 7.

6. The method of claim 5 , wherein the filter is at least one of a Butterworth filter, a Type-I Chebyshev filter, and a Type-II Chebyshev filter.

7. The method of claim 5 , further comprising:

selecting a frequency band of the filter to retain a maximum energy and viable information in the drift removed signal, wherein the frequency band ranges from 0.04 hertz (Hz) to 30 Hz.

8. The method of claim 1 , wherein performing the motion artifact correction on the filtered signal further comprises:

applying a moving average filter (MAF) on the plurality of wavelet coefficients;

obtaining smoothed coefficients, based on applying the MAF on the plurality of wavelet coefficients; and

reconstructing the corrected signal from the smoothed coefficients using an inverse continuous wavelet transform (ICWT).

9. The method of claim 8 , wherein identifying the motion artifact corrupted coefficients comprises:

identifying the motion artifact corrupted coefficients based on obtaining a wavelet scalogram from the wavelet coefficients.

10. The method of claim 8 , wherein the intensity is reduced using a mean and a standard deviation of non-corrupted wavelet coefficients.

11. A system for processing a photoplethysmogram (PPG) signal, the system comprising:

a memory configured to store instructions; and

a processor configured to execute the instructions to:

remove a baseline drift from the PPG signal;

obtain a drift removed signal, based on removing the baseline drift from the PPG signal;

filter the drift removed signal;

obtain a filtered signal, based on filtering the drift removed signal;

perform motion artifact correction on the filtered signal by

obtaining a plurality of wavelet coefficients by applying a continuous wavelet transform (CWT) to the filtered signal;

identifying motion artifact corrupted coefficients from the plurality of wavelet coefficients, as outliers, based on an energy percentage of the plurality of wavelet coefficients being greater than a threshold for a same translation parameter and more than two scaling parameters; and

updating the motion artifact corrupted coefficients by reducing an intensity of the identified motion artifact corrupted coefficients; and

obtain a corrected signal, based on the updated motion artifact corrupted coefficients.

12. The system of claim 11 , wherein the processor is further configured to:

measure at least one of a blood pressure, a heart rate, an oxygen saturation, a glucose level, and a mental stress level based on the corrected signal.

13. The system of claim 11 , wherein the processor is configured to:

perform a multi-level decomposition on the PPG signal using a discrete wavelet transform (DWT);

obtain approximation coefficients, based on performing the multi-level decomposition on the PPG signal;

obtain a drift signal from the approximation coefficients at a pre-defined level;

remove the drift signal from the PPG signal; and

obtain the drift removed signal, based on removing the drift signal from the PPG signal.

14. The system of claim 13 , wherein the multi-level decomposition is performed by using a Daubechies 4 (db4) mother wavelet.

15. The system of claim 11 , wherein filtering of the drift removed signal is performed by use of a filter of order n, and

wherein the order of the filter n is 2, 3, 5, or 7.

16. The system of claim 15 , wherein the filter is at least one of a Butterworth filter, a Type-I Chebyshev filter, and a Type-II Chebyshev filter.

17. The system as claimed in claim 15 , wherein the processor is further configured to:

select a frequency band of the filter to retain a maximum energy and viable information in the drift removed signal, and wherein the frequency band ranges from 0.04 hertz (Hz) to 30 Hz.

18. The system of claim 11 , wherein the processor is configured to perform motion artifact correction on the filtered signal by

applying a moving average filter (MAF) on the plurality of wavelet coefficients;

obtaining smoothed coefficients, based on applying the MAF on the plurality of wavelet coefficients; and

reconstructing the corrected signal from the smoothed coefficients using an inverse continuous wavelet transform (ICWT).

19. The system of claim 18 , wherein the processor is configured to identify the motion artifact corrupted coefficients by obtaining a wavelet scalogram from the wavelet coefficients.

20. The system of claim 18 , wherein the intensity is reduced using a mean and a standard deviation of non-corrupted coefficients.