CLIAMS:1. A method for noise cleaning of photoplethysmogram (PPG) signals for estimating blood pressure (BP) of a user; said method comprising:

a. extracting photoplethysmogram signals from the user using an image capturing device (202) coupled with a mobile communication device (204);
b. up sampling the extracted photoplethysmogram signals using an up sampling module (206);
c. filtering the up sampled photoplethysmogram signals using a filtering module (208);
d. removing uneven baseline drift of each cycle of the up sampled and filtered photoplethysmogram signals using a baseline drift removal module (210);
e. removing outlier cycles of the photoplethysmogram signals by k-means clustering using an outlier removing module (212);
f. modeling remaining cycles of the photoplethysmogram signals after removing outlier cycles with a sum of 2 Gaussian functions using a signal modeling module (214); and
g. extracting time domain features from originally extracted and modeled photoplethysmogram signals using a feature extraction module (216) for estimating blood pressure of the user.

2. The method as claimed in claim 1, wherein the photoplethysmogram signals are extracted from user’s peripheral body parts selected from a group comprising finger, ear, toe and forehead.

3. The method as claimed in claim 1, wherein the photoplethysmogram signals are extracted from the the user using a light emitting source attached to the image capturing device (202) coupled with the mobile communication device (204).

4. The method as claimed in claim 1, wherein the photoplethysmogram signals are extracted in Y domain of YCBCR color space of a captured video using the image capturing device (202) coupled with the mobile communication device (204).

5. The method as claimed in claim 1, wherein the image capturing device (202) coupled with the mobile communication device (204) extracts photoplethysmogram signals as a video stream at 30 fps.

6. The method as claimed in claim 1, wherein the photoplethysmogram signals are extracted as a time series data wherein signal value of photoplethysmogram at nth frame is represented by mean value of Y component of the nth frame.

7. The method as claimed in claim 1, wherein the extracted photoplethysmogram signals are up sampled using linear interpolation.

8. The method as claimed in claim 1, wherein the up sampled photoplethysmogram signals are shifted to its zero mean and applied to a 4th order Butterworth band-pass filter having cutoff frequencies of 0.5 Hz and 5 Hz.

9. The method as claimed in claim 1, wherein the uneven baseline drift of each cycle (F) of the up sampled and filtered photoplethysmogram signals of length k is removed by constructing a second vector T forming a line segment of length k, having endpoints of the second vector T same as the endpoints of each cycle F, along with k-2 equal spaced points in between constructed using linear regression, where the vector F1 = F - T representing the modified cycle with zero baseline.

10. The method as claimed in claim 1, wherein the outlier cycles of the photoplethysmogram signals are removed by splitting each cycle of the photoplethysmogram signals into a plurality of rectangular overlapping windows of equal size, identifying fundamental frequency of the plurality of rectangular overlapping windows, calculating absolute difference from ideal time period, indicating high value of the ideal time period as a wrongly detected cycle, removing wrongly detected outliers using k-means clustering.

11. The method as claimed in claim 1, wherein the time domain features including systolic time; diastolic time; pulse-width at; 33% (B33); 75% (B75) of pulse height; total pulse width of the original signal, along with Gaussian RMS width; C1; C2 of the fitted Gaussian curves; and mode parameters b1 and b2 are extracted from originally extracted and modeled photoplethysmogram signals for estimating blood pressure of the user using machine learning techniques.

12. A system (200) for noise cleaning of photoplethysmogram (PPG) signals for estimating blood pressure (BP) of a user; said system (200) comprising:

a. an image capturing device (202) coupled with a mobile communication device (204), adapted for extracting photoplethysmogram signals from the user;
b. an up sampling module (206), adapted for up sampling the extracted photoplethysmogram signals;
c. a filtering module (208), adapted for filtering the up sampled photoplethysmogram signals;
d. a baseline drift removal module (210), adapted for removing uneven baseline drift of each cycle of the up sampled and filtered photoplethysmogram signals;
e. an outlier removing module (212), adapted for removing outlier cycles of the photoplethysmogram signals by k-means clustering;
f. a signal modeling module (214), adapted for modeling remaining cycles of the photoplethysmogram signals after removing outlier cycles with a sum of 2 Gaussian functions; and
g. a feature extraction module (216), adapted for extracting time domain features from originally extracted and modeled photoplethysmogram signals for estimating blood pressure of the user.

13. The method as claimed in claim 1, wherein the image capturing device (202) coupled with a mobile communication device (204) is adapted to extract photoplethysmogram signals from user’s peripheral body parts selected from a group comprising finger, ear, toe and forehead.

14. The system (200) as claimed in claim 12, wherein the image capturing device (202) coupled with the mobile communication device (204) is having a light emitting source for extracting photoplethysmogram signals.

15. The system (200) as claimed in claim 12, wherein the image capturing device (202) coupled with the mobile communication device (204) is extracting the photoplethysmogram signals in Y domain of YCBCR color space of a captured video.

16. The system (200) as claimed in claim 12, wherein the image capturing device (202) coupled with the mobile communication device (204) extracts photoplethysmogram signals as a video stream at 30 fps.

17. The system (200) as claimed in claim 12, wherein the photoplethysmogram signals are extracted as a time series data wherein signal value of photoplethysmogram at nth frame is represented by mean value of Y component of the nth frame.

18. The system (200) as claimed in claim 12, wherein the extracted photoplethysmogram signals are up sampled using linear interpolation.

19. The system (200) as claimed in claim 12, wherein the up sampled photoplethysmogram signals are shifted to its zero mean and applied to a 4th order Butterworth band-pass filter having cutoff frequencies of 0.5 Hz and 5 Hz.

20. The system (200) as claimed in claim 12, wherein the uneven baseline drift of each cycle (F) of the up sampled and filtered photoplethysmogram signals of length k is removed by constructing a second vector T forming a line segment of length k, having endpoints of the second vector T same as the endpoints of each cycle F, along with k-2 equal spaced points in between constructed using linear regression, where the vector F1 = F - T representing the modified cycle with zero baseline.

21. The system (200) as claimed in claim 12, wherein the outlier cycles of the photoplethysmogram signals are removed by splitting each cycle of the photoplethysmogram signals into a plurality of rectangular overlapping windows of equal size, identifying fundamental frequency of the plurality of rectangular overlapping windows, calculating absolute difference from ideal time period, indicating high value of the ideal time period as a wrongly detected cycle, removing wrongly detected outliers using k-means clustering.

22. The system (200) as claimed in claim 12, wherein the time domain features including systolic time; diastolic time; pulse-width at; 33% (B33); 75% (B75) of pulse height; total pulse width of the original signal, along with Gaussian RMS width; C1; C2 of the fitted Gaussian curves; and mode parameters b1 and b2 are extracted from originally extracted and modeled photoplethysmogram signals for estimating blood pressure of the user using machine learning techniques.
,TagSPECI:As Attached