Methods and systems for extracting venous pulsation and respiratory information from photoplethysmographs
View Patent ↗A system and method for separating a venous component and an arterial component from a red signal and an infrared signal of a PPG sensor is provided. The method uses the second order statistics of venous and arterial signals to separate the venous and arterial signals. After reliable separation of the venous and the arterial component signals, the component signals can be used for different purposes. In a preferred embodiment, the respiratory signal, pattern, and rate are extracted from the separated venous component and a reliable “ratio of ratios” is extracted for SpO 2 using only the arterial component of the PPG signals. The disclosed embodiments enable real-time continuous monitoring of respiration pattern/rate and site-independent arterial oxygen saturation.
1. A method for extracting arterial, venous, blood oxygen, and respiratory information from a set of photoplethysmographic signals comprising the steps of:
preprocessing the set of photoplethysmographic signals;
locating a venous signal and an arterial signal in each photoplethysmographic signal of the set of photoplethysmographic signals, using second order blind identification;
extracting the arterial signal and the venous signal from each photoplethysmographic signal of the set of photoplethysmographic signals;
extracting a respiratory signal using the venous signal and the arterial signal; and,
extracting a blood oxygen saturation from the arterial signal;
wherein the set of photoplethysmographic signals comprises a red signal and an infrared signal;
wherein the red signal comprises a red venous signal and a red arterial signal; and,
wherein the step of extracting a respiratory signal further comprises the steps of:
determining a red signal derivative of the red signal;
determining a red venous signal derivative for the red venous signal;
generating a time domain window signal from the red signal derivative and the red venous signal derivative;
generating a windowed signal from the time domain window signal and the red venous signal derivative;
determining a maximum amplitude from the windowed signal;
determining a venous return variation from the maximum amplitude;
interpolating the venous return variation; and,
filtering the venous return variation to generate the respiratory signal.
2. The method of claim 1 , further comprising the steps of:
computing an autocorrelation function of the respiratory signal;
determining a peak of the autocorrelation function; and,
calculating a respiratory rate using the peak.
3. The method of claim 1 , wherein the step of extracting the arterial signal further comprises the steps of:
determining a venous reference signal from the set of photoplethysmographic signals;
filtering the arterial signal from each photoplethysmographic signal of the set of photoplethysmographic signals, using the venous reference signal; and,
filtering the venous signal from each photoplethysmographic signal of the set of photoplethysmographic signals, using the venous reference signal.
4. The method of claim 3 , further comprising the steps of:
extracting a fundamental frequency from the arterial signal; and,
filtering the venous signal using the fundamental frequency.
5. The method of claim 1 , wherein the step of preprocessing further comprises the step of filtering the set of photoplethysmographic signals with a low pass filter.
6. The method of claim 1 , wherein the step of locating further comprises the steps of:
whitening the set of photoplethysmographic signals; and,
computing an approximate joint diagonalization of the set of photoplethysmographic signals.
7. A method for extracting a respiratory signal and a blood oxygen level from a red photoplethysmographic signal and an infrared photoplethysmographic signal comprising the steps of:
receiving the red photoplethysmographic signal and the infrared photoplethysmographic signal;
filtering the red photoplethysmographic signal and the infrared photoplethysmographic signal;
separating the red photoplethysmographic signal into a first signal component and a second signal component;
separating the infrared photoplethysmographic signal into a third signal component and a fourth signal component;
determining a red venous signal from the first signal component and a red arterial signal from the second signal component;
determining an infrared venous signal from the third signal component and an infrared arterial signal component from the fourth signal component;
extracting the red arterial signal and the red venous signal from the red photoplethysmographic signal;
extracting the infrared arterial signal and the infrared venous signal from the infrared photoplethysmographic signal;
extracting the respiratory signal using the red photoplethysmographic signal and the red venous signal; and,
extracting the blood oxygen signal using the red arterial signal and the infrared arterial signal;
wherein the step of extracting the respiratory signal further comprises the steps of:
determining a red signal derivative of the red photoplethysmographic signal;
determining a red venous signal derivative of the red venous signal;
generating a time domain window signal from the red signal derivative and the red venous signal derivative;
generating a windowed signal from the time domain signal and the red venous signal derivative;
determining a maximum amplitude from the windowed signal;
determining a venous return variation from the maximum amplitude;
interpolating the venous return variation; and,
filtering the venous return variation to generate the respiratory signal.
8. The method of claim 7 , further comprising the steps of:
computing an autocorrelation function of the respiratory signal;
determining a peak of the autocorrelation function; and,
calculating a respiratory rate using the peak.
9. The method of claim 7 , wherein the step of extracting the red arterial signal and the red venous signal further comprises the steps of:
determining a venous reference signal from the red photoplethysmographic signal and the infrared photoplethysmographic signal;
filtering the red arterial signal from the red photoplethysmographic signal using the venous reference signal; and,
filtering the red venous signal from the red photoplethysmographic signal using the venous reference signal.
10. The method of claim 9 , wherein the step of extracting the infrared arterial signal and the infrared venous signal further comprises the steps of:
determining the venous reference signal from the red photoplethysmographic signal and the infrared photoplethysmographic signal;
filtering the infrared arterial signal from the infrared photoplethysmographic signal using the venous reference signal; and,
filtering the infrared venous signal from the infrared photoplethysmographic signal using the venous reference signal.
11. The method of claim 10 , further comprising the steps of:
extracting a fundamental frequency from the infrared arterial signal;
filtering the red venous signal using the fundamental frequency; and
filtering the infrared venous signal using the fundamental frequency.
12. A system for extracting arterial, venous, blood oxygen, and respiratory information from a set of photoplethysmographic signals comprising:
an extractor;
a set of photoplethysmographic sensors outputting a set of photoplethysmographic signals, each photoplethysmographic sensor connected to the extractor;
a computing device connected to the extractor and programmed to:
preprocess the set of photoplethysmographic signals;
locate a venous signal and an arterial signal in each photoplethysmographic signal of the set of photoplethysmographic signals using second order blind identification;
extract the arterial signal and the venous signal from each photoplethysmographic signal of the set of photoplethysmographic signals;
extract a respiratory signal using the venous signal and the arterial signal; and,
extract a blood oxygen saturation from the arterial signal;
wherein each photoplethysmographic sensor comprises a red LED and an infrared LED;
wherein the set of photoplethysmographic signals comprises a red signal and an infrared signal;
wherein the red signal further comprises a red venous signal and a red arterial signal; and,
wherein the computing device is further programmed to:
receive the red signal from the red LED;
receive the infrared signal from the infrared LED;
determine a red signal derivative of the red signal;
determine a red venous signal derivative for the red venous signal;
generate a time domain window signal from the red signal derivative and the red venous signal derivative;
generate a windowed signal from the time domain signal and the red venous signal derivative;
determine a maximum amplitude from the windowed signal;
determine a venous return variation from the maximum amplitude;
interpolate the venous return variation; and,
filter the venous return variation to generate the respiratory signal.
13. The system of claim 12 , wherein the computing device is further programmed to:
compute an autocorrelation function of the respiratory signal;
determine a peak of the autocorrelation function; and,
calculate a respiratory rate using the peak.
14. The system of claim 12 , wherein the computing device is further programmed to:
determine a venous reference signal from the set of photoplethysmographic signals;
filter the arterial signal from each photoplethysmographic signal of the set of photoplethysmographic signals using the venous reference signal; and,
filter the venous signal from each photoplethysmographic signal of the set of photoplethysmographic signals using the venous reference signal.
15. The system of claim 14 , wherein the computing device is further programmed to:
extract a fundamental frequency from the arterial signal; and,
filter the venous signal using the fundamental frequency.
16. The system of claim 12 , wherein the extractor further comprises a set of filters, and wherein the computing device is further programmed to filter the set of photoplethysmographic signals with the set of filters.
17. The system of claim 12 , wherein the computing device is further programmed to:
whiten the set of photoplethysmographic signals; and,
compute an approximate joint diagonalization of the set of photoplethysmographic signals.